[PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi,

QDEV_PROP_PTR is marked in multiple places as "FIXME/TODO/remove
me". In most cases, it can be easily replaced with QDEV_PROP_LINK when
the pointer points to an Object.

There are a few places where such substitution isn't possible. For
those places, it seems reasonable to use a specific setter method
instead, and keep the user_creatable = false. In other places,
proper usage of qdev or other facilies is the solution.

The serial code wasn't converted to qdev, which makes it a bit more
archaic to deal with. Let's convert it first, so we can more easily
embed it from other devices, and re-export some properties and drop
QDEV_PROP_PTR usage.

v4: (after Peter & Philippe reviews)
 - replaced "self" variable names with abbreviations
 - split "mips: inline serial_init()"
 - new patches: "mips: use sysbus_mmio_get_region() instead of internal
   fields", "leon3: use qdev gpio facilities for the PIL", "qdev: use
   g_strcmp0() instead of open-coding it", "qdev/qom: remove some TODO
   limitations now that PROP_PTR is gone"
 - dropped patches: "sparc: move PIL irq handling to cpu.c" & "serial:
   add "instance-id" property"
 - various comments / commit message tweaks
 - added r-b tags

v3:
 - introduce SerialMM and SerialIO sysbus devices
 - remove serial_mm_connect introduced in v2
 - replace "base" property introduced in v2, use "instance-id" for
   vmstate purpose only
 - add a few preliminary clean-ups

v2:
 - qom-ify serial
 - embed the serial from sm501, and expose a "chardev" property
 - add "leon3: use qemu_irq framework instead of callback as property"
 - add "sparc: move PIL irq handling to cpu.c"
 - add "cris: improve passing PIC interrupt vector to the CPU"
 - misc comment/todo changes, add r-b tags

Marc-André Lureau (37):
  qdev: remove unused qdev_prop_int64
  sysbus: remove unused sysbus_try_create*
  sysbus: remove outdated comment
  chardev: generate an internal id when none given
  serial-pci-multi: factor out multi_serial_get_port_count()
  serial: initial qom-ification
  serial: register vmsd with DeviceClass
  serial: add "chardev" property
  serial: add "baudbase" property
  serial: realize the serial device
  serial: replace serial_exit_core() with unrealize
  serial: start making SerialMM a sysbus device
  serial-mm: add "regshift" property
  serial-mm: add endianness property
  serial-mm: use sysbus facilities
  serial: make SerialIO a sysbus device
  mips: inline serial_init()
  mips: baudbase is 115200 by default
  mips: use sysbus_add_io()
  mips: use sysbus_mmio_get_region() instead of internal fields
  sm501: make SerialMM a child, export chardev property
  vmmouse: replace PROP_PTR with PROP_LINK
  lance: replace PROP_PTR with PROP_LINK
  etraxfs: remove PROP_PTR usage
  dp8393x: replace PROP_PTR with PROP_LINK
  leon3: use qemu_irq framework instead of callback as property
  leon3: use qdev gpio facilities for the PIL
  qdev: use g_strcmp0() instead of open-coding it
  RFC: mips/cps: fix setting saar property
  cris: improve passing PIC interrupt vector to the CPU
  smbus-eeprom: remove PROP_PTR
  omap-intc: remove PROP_PTR
  omap-i2c: remove PROP_PTR
  omap-gpio: remove PROP_PTR
  qdev: remove PROP_MEMORY_REGION
  qdev: remove QDEV_PROP_PTR
  qdev/qom: remove some TODO limitations now that PROP_PTR is gone

 chardev/char.c               |  32 +++++--
 hw/arm/omap1.c               |   8 +-
 hw/arm/omap2.c               |  25 +++---
 hw/char/omap_uart.c          |   2 +-
 hw/char/serial-isa.c         |  12 ++-
 hw/char/serial-pci-multi.c   |  55 +++++++-----
 hw/char/serial-pci.c         |  18 +++-
 hw/char/serial.c             | 170 ++++++++++++++++++++++++++++-------
 hw/core/qdev-properties.c    |  50 -----------
 hw/core/qdev.c               |  15 +---
 hw/core/sysbus.c             |  32 -------
 hw/cris/axis_dev88.c         |   4 -
 hw/display/sm501.c           |  33 +++++--
 hw/dma/sparc32_dma.c         |   2 +-
 hw/gpio/omap_gpio.c          |  42 ++++-----
 hw/i2c/omap_i2c.c            |  19 ++--
 hw/i2c/smbus_eeprom.c        |  17 ++--
 hw/i386/pc.c                 |   7 +-
 hw/i386/vmmouse.c            |   8 +-
 hw/input/pckbd.c             |   8 +-
 hw/intc/etraxfs_pic.c        |  26 +-----
 hw/intc/grlib_irqmp.c        |  35 +-------
 hw/intc/omap_intc.c          |  17 ++--
 hw/m68k/q800.c               |   3 +-
 hw/mips/boston.c             |   2 +-
 hw/mips/cps.c                |   2 +-
 hw/mips/mips_jazz.c          |   3 +-
 hw/mips/mips_malta.c         |   2 +-
 hw/mips/mips_mipssim.c       |  14 ++-
 hw/net/dp8393x.c             |   7 +-
 hw/net/etraxfs_eth.c         |  44 ++++++---
 hw/net/lance.c               |   5 +-
 hw/net/pcnet-pci.c           |   2 +-
 hw/net/pcnet.h               |   2 +-
 hw/sh4/r2d.c                 |   2 +-
 hw/sparc/leon3.c             |  15 +++-
 include/hw/arm/omap.h        |  52 +++++++++++
 include/hw/char/serial.h     |  43 ++++++---
 include/hw/cris/etraxfs.h    |  20 +----
 include/hw/input/i8042.h     |   4 +-
 include/hw/qdev-properties.h |  27 ------
 include/hw/sysbus.h          |  13 +--
 include/qemu/id.h            |   1 +
 qom/qom-qmp-cmds.c           |  10 ---
 target/cris/cpu.c            |   8 ++
 target/cris/cpu.h            |   1 +
 util/id.c                    |   1 +
 47 files changed, 497 insertions(+), 423 deletions(-)

-- 
2.24.0

[PATCH v4 01/37] qdev: remove unused qdev_prop_int64

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/core/qdev-properties.c    | 32 --------------------------------
 include/hw/qdev-properties.h |  3 ---
 2 files changed, 35 deletions(-)

diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c
index ac28890e5a..be4cb01f0b 100644
--- a/hw/core/qdev-properties.c
+++ b/hw/core/qdev-properties.c
@@ -409,31 +409,6 @@ static void set_uint64(Object *obj, Visitor *v, const char *name,
     visit_type_uint64(v, name, ptr, errp);
 }
 
-static void get_int64(Object *obj, Visitor *v, const char *name,
-                      void *opaque, Error **errp)
-{
-    DeviceState *dev = DEVICE(obj);
-    Property *prop = opaque;
-    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
-
-    visit_type_int64(v, name, ptr, errp);
-}
-
-static void set_int64(Object *obj, Visitor *v, const char *name,
-                      void *opaque, Error **errp)
-{
-    DeviceState *dev = DEVICE(obj);
-    Property *prop = opaque;
-    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
-
-    if (dev->realized) {
-        qdev_prop_set_after_realize(dev, name, errp);
-        return;
-    }
-
-    visit_type_int64(v, name, ptr, errp);
-}
-
 const PropertyInfo qdev_prop_uint64 = {
     .name  = "uint64",
     .get   = get_uint64,
@@ -441,13 +416,6 @@ const PropertyInfo qdev_prop_uint64 = {
     .set_default_value = set_default_value_uint,
 };
 
-const PropertyInfo qdev_prop_int64 = {
-    .name  = "int64",
-    .get   = get_int64,
-    .set   = set_int64,
-    .set_default_value = set_default_value_int,
-};
-
 /* --- string --- */
 
 static void release_string(Object *obj, const char *name, void *opaque)
diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h
index c6a8cb5516..690ff07ae2 100644
--- a/include/hw/qdev-properties.h
+++ b/include/hw/qdev-properties.h
@@ -13,7 +13,6 @@ extern const PropertyInfo qdev_prop_uint16;
 extern const PropertyInfo qdev_prop_uint32;
 extern const PropertyInfo qdev_prop_int32;
 extern const PropertyInfo qdev_prop_uint64;
-extern const PropertyInfo qdev_prop_int64;
 extern const PropertyInfo qdev_prop_size;
 extern const PropertyInfo qdev_prop_string;
 extern const PropertyInfo qdev_prop_chr;
@@ -164,8 +163,6 @@ extern const PropertyInfo qdev_prop_pcie_link_width;
     DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int32, int32_t)
 #define DEFINE_PROP_UINT64(_n, _s, _f, _d)                      \
     DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_uint64, uint64_t)
-#define DEFINE_PROP_INT64(_n, _s, _f, _d)                      \
-    DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int64, int64_t)
 #define DEFINE_PROP_SIZE(_n, _s, _f, _d)                       \
     DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_size, uint64_t)
 #define DEFINE_PROP_PCI_DEVFN(_n, _s, _f, _d)                   \
-- 
2.24.0

[PATCH v4 02/37] sysbus: remove unused sysbus_try_create*

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/core/sysbus.c    | 32 --------------------------------
 hw/i386/pc.c        |  1 -
 include/hw/sysbus.h |  9 +--------
 3 files changed, 1 insertion(+), 41 deletions(-)

diff --git a/hw/core/sysbus.c b/hw/core/sysbus.c
index 9e69c83aed..08b0311c5f 100644
--- a/hw/core/sysbus.c
+++ b/hw/core/sysbus.c
@@ -250,38 +250,6 @@ DeviceState *sysbus_create_varargs(const char *name,
     return dev;
 }
 
-DeviceState *sysbus_try_create_varargs(const char *name,
-                                       hwaddr addr, ...)
-{
-    DeviceState *dev;
-    SysBusDevice *s;
-    va_list va;
-    qemu_irq irq;
-    int n;
-
-    dev = qdev_try_create(NULL, name);
-    if (!dev) {
-        return NULL;
-    }
-    s = SYS_BUS_DEVICE(dev);
-    qdev_init_nofail(dev);
-    if (addr != (hwaddr)-1) {
-        sysbus_mmio_map(s, 0, addr);
-    }
-    va_start(va, addr);
-    n = 0;
-    while (1) {
-        irq = va_arg(va, qemu_irq);
-        if (!irq) {
-            break;
-        }
-        sysbus_connect_irq(s, n, irq);
-        n++;
-    }
-    va_end(va);
-    return dev;
-}
-
 static void sysbus_dev_print(Monitor *mon, DeviceState *dev, int indent)
 {
     SysBusDevice *s = SYS_BUS_DEVICE(dev);
diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index ac08e63604..49e2259327 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1399,7 +1399,6 @@ void pc_basic_device_init(ISABus *isa_bus, qemu_irq *gsi,
      * when the HPET wants to take over. Thus we have to disable the latter.
      */
     if (!no_hpet && (!kvm_irqchip_in_kernel() || kvm_has_pit_state2())) {
-        /* In order to set property, here not using sysbus_try_create_simple */
         hpet = qdev_try_create(NULL, TYPE_HPET);
         if (hpet) {
             /* For pc-piix-*, hpet's intcap is always IRQ2. For pc-q35-1.7
diff --git a/include/hw/sysbus.h b/include/hw/sysbus.h
index 27e80881da..2eb0484388 100644
--- a/include/hw/sysbus.h
+++ b/include/hw/sysbus.h
@@ -117,8 +117,7 @@ void foreach_dynamic_sysbus_device(FindSysbusDeviceFunc *func, void *opaque);
 /* Legacy helper function for creating devices.  */
 DeviceState *sysbus_create_varargs(const char *name,
                                  hwaddr addr, ...);
-DeviceState *sysbus_try_create_varargs(const char *name,
-                                       hwaddr addr, ...);
+
 static inline DeviceState *sysbus_create_simple(const char *name,
                                               hwaddr addr,
                                               qemu_irq irq)
@@ -126,11 +125,5 @@ static inline DeviceState *sysbus_create_simple(const char *name,
     return sysbus_create_varargs(name, addr, irq, NULL);
 }
 
-static inline DeviceState *sysbus_try_create_simple(const char *name,
-                                                    hwaddr addr,
-                                                    qemu_irq irq)
-{
-    return sysbus_try_create_varargs(name, addr, irq, NULL);
-}
 
 #endif /* HW_SYSBUS_H */
-- 
2.24.0

[PATCH v4 03/37] sysbus: remove outdated comment

[Marc-André Lureau]

The init callback is no more since commit
817a17fc60f44e29a1944b60d32f45ea127f0cf9 ("core/sysbus: remove the
SysBusDeviceClass::init path")

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 include/hw/sysbus.h | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/include/hw/sysbus.h b/include/hw/sysbus.h
index 2eb0484388..c4a1c0adfa 100644
--- a/include/hw/sysbus.h
+++ b/include/hw/sysbus.h
@@ -24,10 +24,6 @@ typedef struct SysBusDevice SysBusDevice;
 
 /**
  * SysBusDeviceClass:
- * @init: Callback function invoked when the #DeviceState.realized property
- * is changed to %true. Deprecated, new types inheriting directly from
- * TYPE_SYS_BUS_DEVICE should use #DeviceClass.realize instead, new leaf
- * types should consult their respective parent type.
  *
  * SysBusDeviceClass is not overriding #DeviceClass.realize, so derived
  * classes overriding it are not required to invoke its implementation.
-- 
2.24.0

[PATCH v4 04/37] chardev: generate an internal id when none given

[Marc-André Lureau]

Internally, qemu may create chardev without ID. Those will not be
looked up with qemu_chr_find(), which prevents using qdev_prop_set_chr().

Use id_generate(), to generate an internal name (prefixed with #), so
no conflict exist with user-named chardev.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 chardev/char.c    | 32 ++++++++++++++++++++++++--------
 include/qemu/id.h |  1 +
 util/id.c         |  1 +
 3 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/chardev/char.c b/chardev/char.c
index 7b6b2cb123..e7e7492b0e 100644
--- a/chardev/char.c
+++ b/chardev/char.c
@@ -37,6 +37,7 @@
 #include "qemu/help_option.h"
 #include "qemu/module.h"
 #include "qemu/option.h"
+#include "qemu/id.h"
 
 #include "chardev/char-mux.h"
 
@@ -944,10 +945,10 @@ void qemu_chr_set_feature(Chardev *chr,
     return set_bit(feature, chr->features);
 }
 
-Chardev *qemu_chardev_new(const char *id, const char *typename,
-                          ChardevBackend *backend,
-                          GMainContext *gcontext,
-                          Error **errp)
+static Chardev *chardev_new(const char *id, const char *typename,
+                            ChardevBackend *backend,
+                            GMainContext *gcontext,
+                            Error **errp)
 {
     Object *obj;
     Chardev *chr = NULL;
@@ -991,6 +992,21 @@ end:
     return chr;
 }
 
+Chardev *qemu_chardev_new(const char *id, const char *typename,
+                          ChardevBackend *backend,
+                          GMainContext *gcontext,
+                          Error **errp)
+{
+    g_autofree char *genid = NULL;
+
+    if (!id) {
+        genid = id_generate(ID_CHR);
+        id = genid;
+    }
+
+    return chardev_new(id, typename, backend, gcontext, errp);
+}
+
 ChardevReturn *qmp_chardev_add(const char *id, ChardevBackend *backend,
                                Error **errp)
 {
@@ -1003,8 +1019,8 @@ ChardevReturn *qmp_chardev_add(const char *id, ChardevBackend *backend,
         return NULL;
     }
 
-    chr = qemu_chardev_new(id, object_class_get_name(OBJECT_CLASS(cc)),
-                           backend, NULL, errp);
+    chr = chardev_new(id, object_class_get_name(OBJECT_CLASS(cc)),
+                      backend, NULL, errp);
     if (!chr) {
         return NULL;
     }
@@ -1061,8 +1077,8 @@ ChardevReturn *qmp_chardev_change(const char *id, ChardevBackend *backend,
         return NULL;
     }
 
-    chr_new = qemu_chardev_new(NULL, object_class_get_name(OBJECT_CLASS(cc)),
-                               backend, chr->gcontext, errp);
+    chr_new = chardev_new(NULL, object_class_get_name(OBJECT_CLASS(cc)),
+                          backend, chr->gcontext, errp);
     if (!chr_new) {
         return NULL;
     }
diff --git a/include/qemu/id.h b/include/qemu/id.h
index 40c70103e4..b55c406e69 100644
--- a/include/qemu/id.h
+++ b/include/qemu/id.h
@@ -4,6 +4,7 @@
 typedef enum IdSubSystems {
     ID_QDEV,
     ID_BLOCK,
+    ID_CHR,
     ID_MAX      /* last element, used as array size */
 } IdSubSystems;
 
diff --git a/util/id.c b/util/id.c
index af1c5f1b81..5addb4460e 100644
--- a/util/id.c
+++ b/util/id.c
@@ -34,6 +34,7 @@ bool id_wellformed(const char *id)
 static const char *const id_subsys_str[ID_MAX] = {
     [ID_QDEV]  = "qdev",
     [ID_BLOCK] = "block",
+    [ID_CHR] = "chr",
 };
 
 /*
-- 
2.24.0

[PATCH v4 05/37] serial-pci-multi: factor out multi_serial_get_port_count()

[Marc-André Lureau]

Common function to be reused in next patch.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/char/serial-pci-multi.c | 35 ++++++++++++++++++-----------------
 1 file changed, 18 insertions(+), 17 deletions(-)

diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
index 5f13b5663b..5c553c30d0 100644
--- a/hw/char/serial-pci-multi.c
+++ b/hw/char/serial-pci-multi.c
@@ -77,33 +77,34 @@ static void multi_serial_irq_mux(void *opaque, int n, int level)
     pci_set_irq(&pci->dev, pending);
 }
 
+static size_t multi_serial_get_port_count(PCIDeviceClass *pc)
+{
+    switch (pc->device_id) {
+    case 0x0003:
+        return 2;
+    case 0x0004:
+        return 4;
+    }
+
+    g_assert_not_reached();
+}
+
+
 static void multi_serial_pci_realize(PCIDevice *dev, Error **errp)
 {
     PCIDeviceClass *pc = PCI_DEVICE_GET_CLASS(dev);
     PCIMultiSerialState *pci = DO_UPCAST(PCIMultiSerialState, dev, dev);
     SerialState *s;
     Error *err = NULL;
-    int i, nr_ports = 0;
-
-    switch (pc->device_id) {
-    case 0x0003:
-        nr_ports = 2;
-        break;
-    case 0x0004:
-        nr_ports = 4;
-        break;
-    }
-    assert(nr_ports > 0);
-    assert(nr_ports <= PCI_SERIAL_MAX_PORTS);
+    size_t i, nports = multi_serial_get_port_count(pc);
 
     pci->dev.config[PCI_CLASS_PROG] = pci->prog_if;
     pci->dev.config[PCI_INTERRUPT_PIN] = 0x01;
-    memory_region_init(&pci->iobar, OBJECT(pci), "multiserial", 8 * nr_ports);
+    memory_region_init(&pci->iobar, OBJECT(pci), "multiserial", 8 * nports);
     pci_register_bar(&pci->dev, 0, PCI_BASE_ADDRESS_SPACE_IO, &pci->iobar);
-    pci->irqs = qemu_allocate_irqs(multi_serial_irq_mux, pci,
-                                   nr_ports);
+    pci->irqs = qemu_allocate_irqs(multi_serial_irq_mux, pci, nports);
 
-    for (i = 0; i < nr_ports; i++) {
+    for (i = 0; i < nports; i++) {
         s = pci->state + i;
         s->baudbase = 115200;
         serial_realize_core(s, &err);
@@ -113,7 +114,7 @@ static void multi_serial_pci_realize(PCIDevice *dev, Error **errp)
             return;
         }
         s->irq = pci->irqs[i];
-        pci->name[i] = g_strdup_printf("uart #%d", i + 1);
+        pci->name[i] = g_strdup_printf("uart #%zu", i + 1);
         memory_region_init_io(&s->io, OBJECT(pci), &serial_io_ops, s,
                               pci->name[i], 8);
         memory_region_add_subregion(&pci->iobar, 8 * i, &s->io);
-- 
2.24.0

[PATCH v4 06/37] serial: initial qom-ification

[Marc-André Lureau]

Make SerialState a device (the following patches will introduce IO/MM
sysbus serial devices)

None of the serial_{,mm}_init() callers actually free the returned
value (even if they did, it would be quite harmless), so we can change
the object allocation at will.

However, the devices that embed SerialState must now have their field
QOM-initialized manually (isa, pci, pci-multi).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial-isa.c       |  9 +++++++++
 hw/char/serial-pci-multi.c | 15 +++++++++++++++
 hw/char/serial-pci.c       | 13 ++++++++++++-
 hw/char/serial.c           | 33 +++++++++++++++++++++++++++------
 include/hw/char/serial.h   |  7 ++++++-
 5 files changed, 69 insertions(+), 8 deletions(-)

diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
index 9e31c51bb6..9a5928b3ee 100644
--- a/hw/char/serial-isa.c
+++ b/hw/char/serial-isa.c
@@ -111,10 +111,19 @@ static void serial_isa_class_initfn(ObjectClass *klass, void *data)
     set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
 }
 
+static void serial_isa_initfn(Object *o)
+{
+    ISASerialState *self = ISA_SERIAL(o);
+
+    object_initialize_child(o, "serial", &self->state, sizeof(self->state),
+                            TYPE_SERIAL, &error_abort, NULL);
+}
+
 static const TypeInfo serial_isa_info = {
     .name          = TYPE_ISA_SERIAL,
     .parent        = TYPE_ISA_DEVICE,
     .instance_size = sizeof(ISASerialState),
+    .instance_init = serial_isa_initfn,
     .class_init    = serial_isa_class_initfn,
 };
 
diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
index 5c553c30d0..edfbfdca9e 100644
--- a/hw/char/serial-pci-multi.c
+++ b/hw/char/serial-pci-multi.c
@@ -181,10 +181,24 @@ static void multi_4x_serial_pci_class_initfn(ObjectClass *klass, void *data)
     set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
 }
 
+static void multi_serial_init(Object *o)
+{
+    PCIDevice *dev = PCI_DEVICE(o);
+    PCIMultiSerialState *pms = DO_UPCAST(PCIMultiSerialState, dev, dev);
+    size_t i, nports = multi_serial_get_port_count(PCI_DEVICE_GET_CLASS(dev));
+
+    for (i = 0; i < nports; i++) {
+        object_initialize_child(o, "serial[*]", &pms->state[i],
+                                sizeof(pms->state[i]),
+                                TYPE_SERIAL, &error_abort, NULL);
+    }
+}
+
 static const TypeInfo multi_2x_serial_pci_info = {
     .name          = "pci-serial-2x",
     .parent        = TYPE_PCI_DEVICE,
     .instance_size = sizeof(PCIMultiSerialState),
+    .instance_init = multi_serial_init,
     .class_init    = multi_2x_serial_pci_class_initfn,
     .interfaces = (InterfaceInfo[]) {
         { INTERFACE_CONVENTIONAL_PCI_DEVICE },
@@ -196,6 +210,7 @@ static const TypeInfo multi_4x_serial_pci_info = {
     .name          = "pci-serial-4x",
     .parent        = TYPE_PCI_DEVICE,
     .instance_size = sizeof(PCIMultiSerialState),
+    .instance_init = multi_serial_init,
     .class_init    = multi_4x_serial_pci_class_initfn,
     .interfaces = (InterfaceInfo[]) {
         { INTERFACE_CONVENTIONAL_PCI_DEVICE },
diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
index cb9b76e22b..f99b6c19e0 100644
--- a/hw/char/serial-pci.c
+++ b/hw/char/serial-pci.c
@@ -40,6 +40,8 @@ typedef struct PCISerialState {
     uint8_t prog_if;
 } PCISerialState;
 
+#define TYPE_PCI_SERIAL "pci-serial"
+#define PCI_SERIAL(s) OBJECT_CHECK(PCISerialState, (s), TYPE_PCI_SERIAL)
 
 static void serial_pci_realize(PCIDevice *dev, Error **errp)
 {
@@ -103,10 +105,19 @@ static void serial_pci_class_initfn(ObjectClass *klass, void *data)
     set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
 }
 
+static void serial_pci_init(Object *o)
+{
+    PCISerialState *ps = PCI_SERIAL(o);
+
+    object_initialize_child(o, "serial", &ps->state, sizeof(ps->state),
+                            TYPE_SERIAL, &error_abort, NULL);
+}
+
 static const TypeInfo serial_pci_info = {
-    .name          = "pci-serial",
+    .name          = TYPE_PCI_SERIAL,
     .parent        = TYPE_PCI_DEVICE,
     .instance_size = sizeof(PCISerialState),
+    .instance_init = serial_pci_init,
     .class_init    = serial_pci_class_initfn,
     .interfaces = (InterfaceInfo[]) {
         { INTERFACE_CONVENTIONAL_PCI_DEVICE },
diff --git a/hw/char/serial.c b/hw/char/serial.c
index b4aa250950..e0a5bec290 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -983,9 +983,8 @@ const MemoryRegionOps serial_io_ops = {
 SerialState *serial_init(int base, qemu_irq irq, int baudbase,
                          Chardev *chr, MemoryRegion *system_io)
 {
-    SerialState *s;
-
-    s = g_malloc0(sizeof(SerialState));
+    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
+    SerialState *s = SERIAL(dev);
 
     s->irq = irq;
     s->baudbase = baudbase;
@@ -993,6 +992,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     serial_realize_core(s, &error_fatal);
 
     vmstate_register(NULL, base, &vmstate_serial, s);
+    qdev_init_nofail(dev);
 
     memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
     memory_region_add_subregion(system_io, base, &s->io);
@@ -1000,6 +1000,20 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     return s;
 }
 
+static void serial_class_init(ObjectClass *klass, void *data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->user_creatable = false;
+}
+
+static const TypeInfo serial_info = {
+    .name = TYPE_SERIAL,
+    .parent = TYPE_DEVICE,
+    .instance_size = sizeof(SerialState),
+    .class_init = serial_class_init,
+};
+
 /* Memory mapped interface */
 static uint64_t serial_mm_read(void *opaque, hwaddr addr,
                                unsigned size)
@@ -1045,9 +1059,8 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
                             qemu_irq irq, int baudbase,
                             Chardev *chr, enum device_endian end)
 {
-    SerialState *s;
-
-    s = g_malloc0(sizeof(SerialState));
+    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
+    SerialState *s = SERIAL(dev);
 
     s->it_shift = it_shift;
     s->irq = irq;
@@ -1056,9 +1069,17 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
 
     serial_realize_core(s, &error_fatal);
     vmstate_register(NULL, base, &vmstate_serial, s);
+    qdev_init_nofail(dev);
 
     memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], s,
                           "serial", 8 << it_shift);
     memory_region_add_subregion(address_space, base, &s->io);
     return s;
 }
+
+static void serial_register_types(void)
+{
+    type_register_static(&serial_info);
+}
+
+type_init(serial_register_types)
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 8be3d8a4f9..180cc7c24e 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -30,10 +30,13 @@
 #include "exec/memory.h"
 #include "qemu/fifo8.h"
 #include "chardev/char.h"
+#include "hw/sysbus.h"
 
 #define UART_FIFO_LENGTH    16      /* 16550A Fifo Length */
 
 typedef struct SerialState {
+    DeviceState parent;
+
     uint16_t divider;
     uint8_t rbr; /* receive register */
     uint8_t thr; /* transmit holding register */
@@ -84,7 +87,9 @@ void serial_realize_core(SerialState *s, Error **errp);
 void serial_exit_core(SerialState *s);
 void serial_set_frequency(SerialState *s, uint32_t frequency);
 
-/* legacy pre qom */
+#define TYPE_SERIAL "serial"
+#define SERIAL(s) OBJECT_CHECK(SerialState, (s), TYPE_SERIAL)
+
 SerialState *serial_init(int base, qemu_irq irq, int baudbase,
                          Chardev *chr, MemoryRegion *system_io);
 SerialState *serial_mm_init(MemoryRegion *address_space,
-- 
2.24.0

[PATCH v4 07/37] serial: register vmsd with DeviceClass

[Marc-André Lureau]

Migration from old to new code works, however the other way fails for
devices that use serial_init/serial_mm_init with "base", used as
instance_id previously.

(with qdev_set_legacy_instance_id, the alias_id is only used in
savevm.c:find_se(), and thus can only be used to match against
"legacy" instance id values. On new code, instance_id is generated
incrementally from 0 with calculate_new_instance_id(), based on
"qdev-path/vmsd-name")

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index e0a5bec290..0c7b9ae45a 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -990,8 +990,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     s->baudbase = baudbase;
     qemu_chr_fe_init(&s->chr, chr, &error_abort);
     serial_realize_core(s, &error_fatal);
-
-    vmstate_register(NULL, base, &vmstate_serial, s);
+    qdev_set_legacy_instance_id(dev, base, 2);
     qdev_init_nofail(dev);
 
     memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
@@ -1005,6 +1004,7 @@ static void serial_class_init(ObjectClass *klass, void *data)
     DeviceClass *dc = DEVICE_CLASS(klass);
 
     dc->user_creatable = false;
+    dc->vmsd = &vmstate_serial;
 }
 
 static const TypeInfo serial_info = {
@@ -1068,7 +1068,7 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
     qemu_chr_fe_init(&s->chr, chr, &error_abort);
 
     serial_realize_core(s, &error_fatal);
-    vmstate_register(NULL, base, &vmstate_serial, s);
+    qdev_set_legacy_instance_id(dev, base, 2);
     qdev_init_nofail(dev);
 
     memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], s,
-- 
2.24.0

[PATCH v4 08/37] serial: add "chardev" property

[Marc-André Lureau]

This is more QOM-friendly, callers may set/get the property themself.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/char/serial.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 0c7b9ae45a..83417cb8c4 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -34,6 +34,7 @@
 #include "sysemu/runstate.h"
 #include "qemu/error-report.h"
 #include "trace.h"
+#include "hw/qdev-properties.h"
 
 //#define DEBUG_SERIAL
 
@@ -988,7 +989,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
 
     s->irq = irq;
     s->baudbase = baudbase;
-    qemu_chr_fe_init(&s->chr, chr, &error_abort);
+    qdev_prop_set_chr(dev, "chardev", chr);
     serial_realize_core(s, &error_fatal);
     qdev_set_legacy_instance_id(dev, base, 2);
     qdev_init_nofail(dev);
@@ -999,12 +1000,18 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     return s;
 }
 
-static void serial_class_init(ObjectClass *klass, void *data)
+static Property serial_properties[] = {
+    DEFINE_PROP_CHR("chardev", SerialState, chr),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void serial_class_init(ObjectClass *klass, void* data)
 {
     DeviceClass *dc = DEVICE_CLASS(klass);
 
     dc->user_creatable = false;
     dc->vmsd = &vmstate_serial;
+    dc->props = serial_properties;
 }
 
 static const TypeInfo serial_info = {
@@ -1065,7 +1072,7 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
     s->it_shift = it_shift;
     s->irq = irq;
     s->baudbase = baudbase;
-    qemu_chr_fe_init(&s->chr, chr, &error_abort);
+    qdev_prop_set_chr(dev, "chardev", chr);
 
     serial_realize_core(s, &error_fatal);
     qdev_set_legacy_instance_id(dev, base, 2);
-- 
2.24.0

[PATCH v4 09/37] serial: add "baudbase" property

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/char/serial-isa.c       | 1 -
 hw/char/serial-pci-multi.c | 1 -
 hw/char/serial-pci.c       | 1 -
 hw/char/serial.c           | 5 +++--
 include/hw/char/serial.h   | 2 +-
 5 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
index 9a5928b3ee..2a4c8de1bf 100644
--- a/hw/char/serial-isa.c
+++ b/hw/char/serial-isa.c
@@ -73,7 +73,6 @@ static void serial_isa_realizefn(DeviceState *dev, Error **errp)
     }
     index++;
 
-    s->baudbase = 115200;
     isa_init_irq(isadev, &s->irq, isa->isairq);
     serial_realize_core(s, errp);
     qdev_set_legacy_instance_id(dev, isa->iobase, 3);
diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
index edfbfdca9e..4891f32230 100644
--- a/hw/char/serial-pci-multi.c
+++ b/hw/char/serial-pci-multi.c
@@ -106,7 +106,6 @@ static void multi_serial_pci_realize(PCIDevice *dev, Error **errp)
 
     for (i = 0; i < nports; i++) {
         s = pci->state + i;
-        s->baudbase = 115200;
         serial_realize_core(s, &err);
         if (err != NULL) {
             error_propagate(errp, err);
diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
index f99b6c19e0..db2c17aafd 100644
--- a/hw/char/serial-pci.c
+++ b/hw/char/serial-pci.c
@@ -49,7 +49,6 @@ static void serial_pci_realize(PCIDevice *dev, Error **errp)
     SerialState *s = &pci->state;
     Error *err = NULL;
 
-    s->baudbase = 115200;
     serial_realize_core(s, &err);
     if (err != NULL) {
         error_propagate(errp, err);
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 83417cb8c4..4a598ead21 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -988,7 +988,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     SerialState *s = SERIAL(dev);
 
     s->irq = irq;
-    s->baudbase = baudbase;
+    qdev_prop_set_uint32(dev, "baudbase", baudbase);
     qdev_prop_set_chr(dev, "chardev", chr);
     serial_realize_core(s, &error_fatal);
     qdev_set_legacy_instance_id(dev, base, 2);
@@ -1002,6 +1002,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
 
 static Property serial_properties[] = {
     DEFINE_PROP_CHR("chardev", SerialState, chr),
+    DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -1071,7 +1072,7 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
 
     s->it_shift = it_shift;
     s->irq = irq;
-    s->baudbase = baudbase;
+    qdev_prop_set_uint32(dev, "baudbase", baudbase);
     qdev_prop_set_chr(dev, "chardev", chr);
 
     serial_realize_core(s, &error_fatal);
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 180cc7c24e..3dc618598e 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -58,7 +58,7 @@ typedef struct SerialState {
     CharBackend chr;
     int last_break_enable;
     int it_shift;
-    int baudbase;
+    uint32_t baudbase;
     uint32_t tsr_retry;
     guint watch_tag;
     uint32_t wakeup;
-- 
2.24.0

[PATCH v4 10/37] serial: realize the serial device

[Marc-André Lureau]

Instead of calling serial_realize_core(), use the QDev realize
callback.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial-isa.c       | 2 +-
 hw/char/serial-pci-multi.c | 2 +-
 hw/char/serial-pci.c       | 2 +-
 hw/char/serial.c           | 8 ++++----
 include/hw/char/serial.h   | 1 -
 5 files changed, 7 insertions(+), 8 deletions(-)

diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
index 2a4c8de1bf..db8644551e 100644
--- a/hw/char/serial-isa.c
+++ b/hw/char/serial-isa.c
@@ -74,7 +74,7 @@ static void serial_isa_realizefn(DeviceState *dev, Error **errp)
     index++;
 
     isa_init_irq(isadev, &s->irq, isa->isairq);
-    serial_realize_core(s, errp);
+    object_property_set_bool(OBJECT(s), true, "realized", errp);
     qdev_set_legacy_instance_id(dev, isa->iobase, 3);
 
     memory_region_init_io(&s->io, OBJECT(isa), &serial_io_ops, s, "serial", 8);
diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
index 4891f32230..0e1fdb75d2 100644
--- a/hw/char/serial-pci-multi.c
+++ b/hw/char/serial-pci-multi.c
@@ -106,7 +106,7 @@ static void multi_serial_pci_realize(PCIDevice *dev, Error **errp)
 
     for (i = 0; i < nports; i++) {
         s = pci->state + i;
-        serial_realize_core(s, &err);
+        object_property_set_bool(OBJECT(s), true, "realized", &err);
         if (err != NULL) {
             error_propagate(errp, err);
             multi_serial_pci_exit(dev);
diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
index db2c17aafd..4b6a217365 100644
--- a/hw/char/serial-pci.c
+++ b/hw/char/serial-pci.c
@@ -49,7 +49,7 @@ static void serial_pci_realize(PCIDevice *dev, Error **errp)
     SerialState *s = &pci->state;
     Error *err = NULL;
 
-    serial_realize_core(s, &err);
+    object_property_set_bool(OBJECT(s), true, "realized", &err);
     if (err != NULL) {
         error_propagate(errp, err);
         return;
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 4a598ead21..1f3859eef1 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -934,8 +934,10 @@ static int serial_be_change(void *opaque)
     return 0;
 }
 
-void serial_realize_core(SerialState *s, Error **errp)
+static void serial_realize(DeviceState *dev, Error **errp)
 {
+    SerialState *s = SERIAL(dev);
+
     s->modem_status_poll = timer_new_ns(QEMU_CLOCK_VIRTUAL, (QEMUTimerCB *) serial_update_msl, s);
 
     s->fifo_timeout_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, (QEMUTimerCB *) fifo_timeout_int, s);
@@ -990,7 +992,6 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
     s->irq = irq;
     qdev_prop_set_uint32(dev, "baudbase", baudbase);
     qdev_prop_set_chr(dev, "chardev", chr);
-    serial_realize_core(s, &error_fatal);
     qdev_set_legacy_instance_id(dev, base, 2);
     qdev_init_nofail(dev);
 
@@ -1011,6 +1012,7 @@ static void serial_class_init(ObjectClass *klass, void* data)
     DeviceClass *dc = DEVICE_CLASS(klass);
 
     dc->user_creatable = false;
+    dc->realize = serial_realize;
     dc->vmsd = &vmstate_serial;
     dc->props = serial_properties;
 }
@@ -1074,8 +1076,6 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
     s->irq = irq;
     qdev_prop_set_uint32(dev, "baudbase", baudbase);
     qdev_prop_set_chr(dev, "chardev", chr);
-
-    serial_realize_core(s, &error_fatal);
     qdev_set_legacy_instance_id(dev, base, 2);
     qdev_init_nofail(dev);
 
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 3dc618598e..571aab97c8 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -83,7 +83,6 @@ typedef struct SerialState {
 extern const VMStateDescription vmstate_serial;
 extern const MemoryRegionOps serial_io_ops;
 
-void serial_realize_core(SerialState *s, Error **errp);
 void serial_exit_core(SerialState *s);
 void serial_set_frequency(SerialState *s, uint32_t frequency);
 
-- 
2.24.0

[PATCH v4 11/37] serial: replace serial_exit_core() with unrealize

[Marc-André Lureau]

Instead of calling serial_exit_core() directly, use the QDev unrealize
callback.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/char/serial-pci-multi.c | 2 +-
 hw/char/serial-pci.c       | 2 +-
 hw/char/serial.c           | 5 ++++-
 include/hw/char/serial.h   | 1 -
 4 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
index 0e1fdb75d2..e343a1235c 100644
--- a/hw/char/serial-pci-multi.c
+++ b/hw/char/serial-pci-multi.c
@@ -56,7 +56,7 @@ static void multi_serial_pci_exit(PCIDevice *dev)
 
     for (i = 0; i < pci->ports; i++) {
         s = pci->state + i;
-        serial_exit_core(s);
+        object_property_set_bool(OBJECT(s), false, "realized", NULL);
         memory_region_del_subregion(&pci->iobar, &s->io);
         g_free(pci->name[i]);
     }
diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
index 4b6a217365..b6a73c65a9 100644
--- a/hw/char/serial-pci.c
+++ b/hw/char/serial-pci.c
@@ -68,7 +68,7 @@ static void serial_pci_exit(PCIDevice *dev)
     PCISerialState *pci = DO_UPCAST(PCISerialState, dev, dev);
     SerialState *s = &pci->state;
 
-    serial_exit_core(s);
+    object_property_set_bool(OBJECT(s), false, "realized", NULL);
     qemu_free_irq(s->irq);
 }
 
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 1f3859eef1..2d8471dbbb 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -950,8 +950,10 @@ static void serial_realize(DeviceState *dev, Error **errp)
     serial_reset(s);
 }
 
-void serial_exit_core(SerialState *s)
+static void serial_unrealize(DeviceState *dev, Error **errp)
 {
+    SerialState *s = SERIAL(dev);
+
     qemu_chr_fe_deinit(&s->chr, false);
 
     timer_del(s->modem_status_poll);
@@ -1013,6 +1015,7 @@ static void serial_class_init(ObjectClass *klass, void* data)
 
     dc->user_creatable = false;
     dc->realize = serial_realize;
+    dc->unrealize = serial_unrealize;
     dc->vmsd = &vmstate_serial;
     dc->props = serial_properties;
 }
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 571aab97c8..548944b06a 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -83,7 +83,6 @@ typedef struct SerialState {
 extern const VMStateDescription vmstate_serial;
 extern const MemoryRegionOps serial_io_ops;
 
-void serial_exit_core(SerialState *s);
 void serial_set_frequency(SerialState *s, uint32_t frequency);
 
 #define TYPE_SERIAL "serial"
-- 
2.24.0

[PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Marc-André Lureau]

Memory mapped serial device is in fact a sysbus device. The following
patches will make use of sysbus facilities for resource and
registration. In particular, "serial-mm: use sysbus facilities" will
move internal serial realization to serial_mm_realize callback to
follow qdev best practices.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/omap_uart.c      |  2 +-
 hw/char/serial.c         | 47 ++++++++++++++++++++++++++++------------
 hw/mips/boston.c         |  2 +-
 hw/mips/mips_malta.c     |  2 +-
 include/hw/char/serial.h | 20 ++++++++++++-----
 5 files changed, 51 insertions(+), 22 deletions(-)

diff --git a/hw/char/omap_uart.c b/hw/char/omap_uart.c
index 13e4f43c4c..e8da933378 100644
--- a/hw/char/omap_uart.c
+++ b/hw/char/omap_uart.c
@@ -27,7 +27,7 @@
 struct omap_uart_s {
     MemoryRegion iomem;
     hwaddr base;
-    SerialState *serial; /* TODO */
+    SerialMM *serial; /* TODO */
     struct omap_target_agent_s *ta;
     omap_clk fclk;
     qemu_irq irq;
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 2d8471dbbb..cb9cbd5146 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -1031,16 +1031,16 @@ static const TypeInfo serial_info = {
 static uint64_t serial_mm_read(void *opaque, hwaddr addr,
                                unsigned size)
 {
-    SerialState *s = opaque;
-    return serial_ioport_read(s, addr >> s->it_shift, 1);
+    SerialMM *s = SERIAL_MM(opaque);
+    return serial_ioport_read(&s->serial, addr >> s->it_shift, 1);
 }
 
 static void serial_mm_write(void *opaque, hwaddr addr,
                             uint64_t value, unsigned size)
 {
-    SerialState *s = opaque;
+    SerialMM *s = SERIAL_MM(opaque);
     value &= 255;
-    serial_ioport_write(s, addr >> s->it_shift, value, 1);
+    serial_ioport_write(&s->serial, addr >> s->it_shift, value, 1);
 }
 
 static const MemoryRegionOps serial_mm_ops[3] = {
@@ -1067,30 +1067,49 @@ static const MemoryRegionOps serial_mm_ops[3] = {
     },
 };
 
-SerialState *serial_mm_init(MemoryRegion *address_space,
+SerialMM *serial_mm_init(MemoryRegion *address_space,
                             hwaddr base, int it_shift,
                             qemu_irq irq, int baudbase,
                             Chardev *chr, enum device_endian end)
 {
-    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
-    SerialState *s = SERIAL(dev);
+    SerialMM *smm = SERIAL_MM(qdev_create(NULL, TYPE_SERIAL_MM));
+    SerialState *s = &smm->serial;
 
-    s->it_shift = it_shift;
+    smm->it_shift = it_shift;
     s->irq = irq;
-    qdev_prop_set_uint32(dev, "baudbase", baudbase);
-    qdev_prop_set_chr(dev, "chardev", chr);
-    qdev_set_legacy_instance_id(dev, base, 2);
-    qdev_init_nofail(dev);
+    qdev_prop_set_uint32(DEVICE(s), "baudbase", baudbase);
+    qdev_prop_set_chr(DEVICE(s), "chardev", chr);
+    qdev_set_legacy_instance_id(DEVICE(s), base, 2);
+
+    qdev_init_nofail(DEVICE(s));
+    qdev_init_nofail(DEVICE(smm));
 
-    memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], s,
+    memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], smm,
                           "serial", 8 << it_shift);
     memory_region_add_subregion(address_space, base, &s->io);
-    return s;
+
+    return smm;
+}
+
+static void serial_mm_instance_init(Object *o)
+{
+    SerialMM *smm = SERIAL_MM(o);
+
+    object_initialize_child(o, "serial", &smm->serial, sizeof(smm->serial),
+                            TYPE_SERIAL, &error_abort, NULL);
 }
 
+static const TypeInfo serial_mm_info = {
+    .name = TYPE_SERIAL_MM,
+    .parent = TYPE_SYS_BUS_DEVICE,
+    .instance_init = serial_mm_instance_init,
+    .instance_size = sizeof(SerialMM),
+};
+
 static void serial_register_types(void)
 {
     type_register_static(&serial_info);
+    type_register_static(&serial_mm_info);
 }
 
 type_init(serial_register_types)
diff --git a/hw/mips/boston.c b/hw/mips/boston.c
index ca7d813a52..23fdd5ec6a 100644
--- a/hw/mips/boston.c
+++ b/hw/mips/boston.c
@@ -50,7 +50,7 @@ typedef struct {
 
     MachineState *mach;
     MIPSCPSState cps;
-    SerialState *uart;
+    SerialMM *uart;
 
     CharBackend lcd_display;
     char lcd_content[8];
diff --git a/hw/mips/mips_malta.c b/hw/mips/mips_malta.c
index 92e9ca5bfa..89b5e3123a 100644
--- a/hw/mips/mips_malta.c
+++ b/hw/mips/mips_malta.c
@@ -83,7 +83,7 @@ typedef struct {
     uint32_t i2csel;
     CharBackend display;
     char display_text[9];
-    SerialState *uart;
+    SerialMM *uart;
     bool display_inited;
 } MaltaFPGAState;
 
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 548944b06a..730165347c 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -57,7 +57,6 @@ typedef struct SerialState {
     qemu_irq irq;
     CharBackend chr;
     int last_break_enable;
-    int it_shift;
     uint32_t baudbase;
     uint32_t tsr_retry;
     guint watch_tag;
@@ -80,6 +79,14 @@ typedef struct SerialState {
     MemoryRegion io;
 } SerialState;
 
+typedef struct SerialMM {
+    SysBusDevice parent;
+
+    SerialState serial;
+
+    int it_shift;
+} SerialMM;
+
 extern const VMStateDescription vmstate_serial;
 extern const MemoryRegionOps serial_io_ops;
 
@@ -88,12 +95,15 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
 #define TYPE_SERIAL "serial"
 #define SERIAL(s) OBJECT_CHECK(SerialState, (s), TYPE_SERIAL)
 
+#define TYPE_SERIAL_MM "serial-mm"
+#define SERIAL_MM(s) OBJECT_CHECK(SerialMM, (s), TYPE_SERIAL_MM)
+
 SerialState *serial_init(int base, qemu_irq irq, int baudbase,
                          Chardev *chr, MemoryRegion *system_io);
-SerialState *serial_mm_init(MemoryRegion *address_space,
-                            hwaddr base, int it_shift,
-                            qemu_irq irq, int baudbase,
-                            Chardev *chr, enum device_endian end);
+SerialMM *serial_mm_init(MemoryRegion *address_space,
+                         hwaddr base, int it_shift,
+                         qemu_irq irq, int baudbase,
+                         Chardev *chr, enum device_endian end);
 
 /* serial-isa.c */
 
-- 
2.24.0

[PATCH v4 13/37] serial-mm: add "regshift" property

[Marc-André Lureau]

And a property and rename "it_shift" field to "regshift", as it seems
to be more popular (and I don't know what "it" stands for).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/char/serial.c         | 28 +++++++++++++++++++++++-----
 include/hw/char/serial.h |  4 ++--
 2 files changed, 25 insertions(+), 7 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index cb9cbd5146..225f09f30f 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -1032,7 +1032,7 @@ static uint64_t serial_mm_read(void *opaque, hwaddr addr,
                                unsigned size)
 {
     SerialMM *s = SERIAL_MM(opaque);
-    return serial_ioport_read(&s->serial, addr >> s->it_shift, 1);
+    return serial_ioport_read(&s->serial, addr >> s->regshift, 1);
 }
 
 static void serial_mm_write(void *opaque, hwaddr addr,
@@ -1040,7 +1040,7 @@ static void serial_mm_write(void *opaque, hwaddr addr,
 {
     SerialMM *s = SERIAL_MM(opaque);
     value &= 255;
-    serial_ioport_write(&s->serial, addr >> s->it_shift, value, 1);
+    serial_ioport_write(&s->serial, addr >> s->regshift, value, 1);
 }
 
 static const MemoryRegionOps serial_mm_ops[3] = {
@@ -1068,14 +1068,14 @@ static const MemoryRegionOps serial_mm_ops[3] = {
 };
 
 SerialMM *serial_mm_init(MemoryRegion *address_space,
-                            hwaddr base, int it_shift,
+                            hwaddr base, int regshift,
                             qemu_irq irq, int baudbase,
                             Chardev *chr, enum device_endian end)
 {
     SerialMM *smm = SERIAL_MM(qdev_create(NULL, TYPE_SERIAL_MM));
     SerialState *s = &smm->serial;
 
-    smm->it_shift = it_shift;
+    qdev_prop_set_uint8(DEVICE(smm), "regshift", regshift);
     s->irq = irq;
     qdev_prop_set_uint32(DEVICE(s), "baudbase", baudbase);
     qdev_prop_set_chr(DEVICE(s), "chardev", chr);
@@ -1085,7 +1085,7 @@ SerialMM *serial_mm_init(MemoryRegion *address_space,
     qdev_init_nofail(DEVICE(smm));
 
     memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], smm,
-                          "serial", 8 << it_shift);
+                          "serial", 8 << regshift);
     memory_region_add_subregion(address_space, base, &s->io);
 
     return smm;
@@ -1099,11 +1099,29 @@ static void serial_mm_instance_init(Object *o)
                             TYPE_SERIAL, &error_abort, NULL);
 }
 
+static Property serial_mm_properties[] = {
+    /*
+     * Set the spacing between adjacent memory-mapped UART registers.
+     * Each register will be at (1 << regshift) bytes after the
+     * previous one.
+     */
+    DEFINE_PROP_UINT8("regshift", SerialMM, regshift, 0),
+    DEFINE_PROP_END_OF_LIST(),
+};
+
+static void serial_mm_class_init(ObjectClass *klass, void* data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->props = serial_mm_properties;
+}
+
 static const TypeInfo serial_mm_info = {
     .name = TYPE_SERIAL_MM,
     .parent = TYPE_SYS_BUS_DEVICE,
     .instance_init = serial_mm_instance_init,
     .instance_size = sizeof(SerialMM),
+    .class_init = serial_mm_class_init,
 };
 
 static void serial_register_types(void)
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index 730165347c..eaeda62698 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -84,7 +84,7 @@ typedef struct SerialMM {
 
     SerialState serial;
 
-    int it_shift;
+    uint8_t regshift;
 } SerialMM;
 
 extern const VMStateDescription vmstate_serial;
@@ -101,7 +101,7 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
 SerialState *serial_init(int base, qemu_irq irq, int baudbase,
                          Chardev *chr, MemoryRegion *system_io);
 SerialMM *serial_mm_init(MemoryRegion *address_space,
-                         hwaddr base, int it_shift,
+                         hwaddr base, int regshift,
                          qemu_irq irq, int baudbase,
                          Chardev *chr, enum device_endian end);
 
-- 
2.24.0

[PATCH v4 14/37] serial-mm: add endianness property

[Marc-André Lureau]

Add a qdev property for endianness, so memory region setup can be done
in realize.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/char/serial.c         | 2 ++
 include/hw/char/serial.h | 1 +
 2 files changed, 3 insertions(+)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 225f09f30f..1c16fed4d1 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -1080,6 +1080,7 @@ SerialMM *serial_mm_init(MemoryRegion *address_space,
     qdev_prop_set_uint32(DEVICE(s), "baudbase", baudbase);
     qdev_prop_set_chr(DEVICE(s), "chardev", chr);
     qdev_set_legacy_instance_id(DEVICE(s), base, 2);
+    qdev_prop_set_uint8(DEVICE(self), "endianness", end);
 
     qdev_init_nofail(DEVICE(s));
     qdev_init_nofail(DEVICE(smm));
@@ -1106,6 +1107,7 @@ static Property serial_mm_properties[] = {
      * previous one.
      */
     DEFINE_PROP_UINT8("regshift", SerialMM, regshift, 0),
+    DEFINE_PROP_UINT8("endianness", SerialMM, endianness, DEVICE_NATIVE_ENDIAN),
     DEFINE_PROP_END_OF_LIST(),
 };
 
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index eaeda62698..d586924937 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -85,6 +85,7 @@ typedef struct SerialMM {
     SerialState serial;
 
     uint8_t regshift;
+    uint8_t endianness;
 } SerialMM;
 
 extern const VMStateDescription vmstate_serial;
-- 
2.24.0

[PATCH v4 15/37] serial-mm: use sysbus facilities

[Marc-André Lureau]

Make SerialMM a regular sysbus device, by registering the irq, and the
mmio region. Reexport the internal serial properties.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/char/serial.c | 35 ++++++++++++++++++++++++-----------
 1 file changed, 24 insertions(+), 11 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 1c16fed4d1..e0f1a99e8b 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -1073,21 +1073,18 @@ SerialMM *serial_mm_init(MemoryRegion *address_space,
                             Chardev *chr, enum device_endian end)
 {
     SerialMM *smm = SERIAL_MM(qdev_create(NULL, TYPE_SERIAL_MM));
-    SerialState *s = &smm->serial;
+    MemoryRegion *mr;
 
     qdev_prop_set_uint8(DEVICE(smm), "regshift", regshift);
-    s->irq = irq;
-    qdev_prop_set_uint32(DEVICE(s), "baudbase", baudbase);
-    qdev_prop_set_chr(DEVICE(s), "chardev", chr);
-    qdev_set_legacy_instance_id(DEVICE(s), base, 2);
-    qdev_prop_set_uint8(DEVICE(self), "endianness", end);
-
-    qdev_init_nofail(DEVICE(s));
+    qdev_prop_set_uint32(DEVICE(smm), "baudbase", baudbase);
+    qdev_prop_set_chr(DEVICE(smm), "chardev", chr);
+    qdev_set_legacy_instance_id(DEVICE(smm), base, 2);
+    qdev_prop_set_uint8(DEVICE(smm), "endianness", end);
     qdev_init_nofail(DEVICE(smm));
 
-    memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], smm,
-                          "serial", 8 << regshift);
-    memory_region_add_subregion(address_space, base, &s->io);
+    sysbus_connect_irq(SYS_BUS_DEVICE(smm), 0, irq);
+    mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(smm), 0);
+    memory_region_add_subregion(address_space, base, mr);
 
     return smm;
 }
@@ -1098,6 +1095,8 @@ static void serial_mm_instance_init(Object *o)
 
     object_initialize_child(o, "serial", &smm->serial, sizeof(smm->serial),
                             TYPE_SERIAL, &error_abort, NULL);
+
+    qdev_alias_all_properties(DEVICE(&smm->serial), o);
 }
 
 static Property serial_mm_properties[] = {
@@ -1111,11 +1110,25 @@ static Property serial_mm_properties[] = {
     DEFINE_PROP_END_OF_LIST(),
 };
 
+static void serial_mm_realize(DeviceState *dev, Error **errp)
+{
+    SerialMM *smm = SERIAL_MM(dev);
+    SerialState *s = &smm->serial;
+
+    qdev_init_nofail(DEVICE(s));
+
+    memory_region_init_io(&s->io, NULL, &serial_mm_ops[smm->endianness], smm,
+                          "serial", 8 << smm->regshift);
+    sysbus_init_mmio(SYS_BUS_DEVICE(smm), &s->io);
+    sysbus_init_irq(SYS_BUS_DEVICE(smm), &smm->serial.irq);
+}
+
 static void serial_mm_class_init(ObjectClass *klass, void* data)
 {
     DeviceClass *dc = DEVICE_CLASS(klass);
 
     dc->props = serial_mm_properties;
+    dc->realize = serial_mm_realize;
 }
 
 static const TypeInfo serial_mm_info = {
-- 
2.24.0

[PATCH v4 16/37] serial: make SerialIO a sysbus device

[Marc-André Lureau]

Make serial IO a proper sysbus device, similar to serial MM.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial.c         | 59 ++++++++++++++++++++++++++++++++--------
 include/hw/char/serial.h | 13 +++++++--
 2 files changed, 59 insertions(+), 13 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index e0f1a99e8b..164146ede8 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -985,22 +985,58 @@ const MemoryRegionOps serial_io_ops = {
     .endianness = DEVICE_LITTLE_ENDIAN,
 };
 
-SerialState *serial_init(int base, qemu_irq irq, int baudbase,
-                         Chardev *chr, MemoryRegion *system_io)
+static void serial_io_realize(DeviceState *dev, Error **errp)
 {
-    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
-    SerialState *s = SERIAL(dev);
+    SerialIO *sio = SERIAL_IO(dev);
+    SerialState *s = &sio->serial;
 
-    s->irq = irq;
-    qdev_prop_set_uint32(dev, "baudbase", baudbase);
-    qdev_prop_set_chr(dev, "chardev", chr);
-    qdev_set_legacy_instance_id(dev, base, 2);
-    qdev_init_nofail(dev);
+    qdev_init_nofail(DEVICE(s));
 
     memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
-    memory_region_add_subregion(system_io, base, &s->io);
+    sysbus_init_irq(SYS_BUS_DEVICE(sio), &s->irq);
+}
+
+static void serial_io_class_init(ObjectClass *klass, void* data)
+{
+    DeviceClass *dc = DEVICE_CLASS(klass);
+
+    dc->realize = serial_io_realize;
+    /* No dc->vmsd: class has no migratable state */
+}
+
+static void serial_io_instance_init(Object *o)
+{
+    SerialIO *sio = SERIAL_IO(o);
+
+    object_initialize_child(o, "serial", &sio->serial, sizeof(sio->serial),
+                            TYPE_SERIAL, &error_abort, NULL);
+
+    qdev_alias_all_properties(DEVICE(&sio->serial), o);
+}
+
+
+static const TypeInfo serial_io_info = {
+    .name = TYPE_SERIAL_IO,
+    .parent = TYPE_SYS_BUS_DEVICE,
+    .instance_size = sizeof(SerialIO),
+    .instance_init = serial_io_instance_init,
+    .class_init = serial_io_class_init,
+};
+
+SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
+                         Chardev *chr, MemoryRegion *system_io)
+{
+    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
+
+    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
+    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
+    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
+    qdev_init_nofail(DEVICE(sio));
+
+    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
+    memory_region_add_subregion(system_io, base, &sio->serial.io);
 
-    return s;
+    return sio;
 }
 
 static Property serial_properties[] = {
@@ -1142,6 +1178,7 @@ static const TypeInfo serial_mm_info = {
 static void serial_register_types(void)
 {
     type_register_static(&serial_info);
+    type_register_static(&serial_io_info);
     type_register_static(&serial_mm_info);
 }
 
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index d586924937..d356ba838c 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -88,6 +88,12 @@ typedef struct SerialMM {
     uint8_t endianness;
 } SerialMM;
 
+typedef struct SerialIO {
+    SysBusDevice parent;
+
+    SerialState serial;
+} SerialIO;
+
 extern const VMStateDescription vmstate_serial;
 extern const MemoryRegionOps serial_io_ops;
 
@@ -99,8 +105,11 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
 #define TYPE_SERIAL_MM "serial-mm"
 #define SERIAL_MM(s) OBJECT_CHECK(SerialMM, (s), TYPE_SERIAL_MM)
 
-SerialState *serial_init(int base, qemu_irq irq, int baudbase,
-                         Chardev *chr, MemoryRegion *system_io);
+#define TYPE_SERIAL_IO "serial-io"
+#define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
+
+SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
+                      Chardev *chr, MemoryRegion *system_io);
 SerialMM *serial_mm_init(MemoryRegion *address_space,
                          hwaddr base, int regshift,
                          qemu_irq irq, int baudbase,
-- 
2.24.0

[PATCH v4 17/37] mips: inline serial_init()

[Marc-André Lureau]

The function is specific to mipssim, let's inline it.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial.c         | 16 ----------------
 hw/mips/mips_mipssim.c   | 15 ++++++++++++---
 include/hw/char/serial.h |  2 --
 3 files changed, 12 insertions(+), 21 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 164146ede8..23f0b02516 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
     .class_init = serial_io_class_init,
 };
 
-SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
-                         Chardev *chr, MemoryRegion *system_io)
-{
-    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
-
-    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
-    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
-    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
-    qdev_init_nofail(DEVICE(sio));
-
-    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
-    memory_region_add_subregion(system_io, base, &sio->serial.io);
-
-    return sio;
-}
-
 static Property serial_properties[] = {
     DEFINE_PROP_CHR("chardev", SerialState, chr),
     DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
index 282bbecb24..bfafa4d7e9 100644
--- a/hw/mips/mips_mipssim.c
+++ b/hw/mips/mips_mipssim.c
@@ -40,6 +40,7 @@
 #include "hw/loader.h"
 #include "elf.h"
 #include "hw/sysbus.h"
+#include "hw/qdev-properties.h"
 #include "exec/address-spaces.h"
 #include "qemu/error-report.h"
 #include "sysemu/qtest.h"
@@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
      * A single 16450 sits at offset 0x3f8. It is attached to
      * MIPS CPU INT2, which is interrupt 4.
      */
-    if (serial_hd(0))
-        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
-                    get_system_io());
+    if (serial_hd(0)) {
+        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
+
+        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
+        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
+        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
+        qdev_init_nofail(dev);
+        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
+        memory_region_add_subregion(get_system_io(), 0x3f8,
+                                    &SERIAL_IO(dev)->serial.io);
+    }
 
     if (nd_table[0].used)
         /* MIPSnet uses the MIPS CPU INT0, which is interrupt 2. */
diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
index d356ba838c..535fa23a2b 100644
--- a/include/hw/char/serial.h
+++ b/include/hw/char/serial.h
@@ -108,8 +108,6 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
 #define TYPE_SERIAL_IO "serial-io"
 #define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
 
-SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
-                      Chardev *chr, MemoryRegion *system_io);
 SerialMM *serial_mm_init(MemoryRegion *address_space,
                          hwaddr base, int regshift,
                          qemu_irq irq, int baudbase,
-- 
2.24.0

[PATCH v4 18/37] mips: baudbase is 115200 by default

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/mips/mips_mipssim.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
index bfafa4d7e9..3cd0e6eb33 100644
--- a/hw/mips/mips_mipssim.c
+++ b/hw/mips/mips_mipssim.c
@@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
     if (serial_hd(0)) {
         DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
 
-        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
         qdev_prop_set_chr(dev, "chardev", serial_hd(0));
         qdev_set_legacy_instance_id(dev, 0x3f8, 2);
         qdev_init_nofail(dev);
-- 
2.24.0

[PATCH v4 19/37] mips: use sysbus_add_io()

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/mips/mips_mipssim.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
index 3cd0e6eb33..2c2c7f25b2 100644
--- a/hw/mips/mips_mipssim.c
+++ b/hw/mips/mips_mipssim.c
@@ -227,8 +227,7 @@ mips_mipssim_init(MachineState *machine)
         qdev_set_legacy_instance_id(dev, 0x3f8, 2);
         qdev_init_nofail(dev);
         sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
-        memory_region_add_subregion(get_system_io(), 0x3f8,
-                                    &SERIAL_IO(dev)->serial.io);
+        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->serial.io);
     }
 
     if (nd_table[0].used)
-- 
2.24.0

[PATCH v4 20/37] mips: use sysbus_mmio_get_region() instead of internal fields

[Marc-André Lureau]

Register the memory region with sysbus_init_mmio() and look it up with
sysbus_mmio_get_region() to avoid accessing internal device fields.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/char/serial.c       | 1 +
 hw/mips/mips_mipssim.c | 3 ++-
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 23f0b02516..02c545ff8c 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -993,6 +993,7 @@ static void serial_io_realize(DeviceState *dev, Error **errp)
     qdev_init_nofail(DEVICE(s));
 
     memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
+    sysbus_init_mmio(SYS_BUS_DEVICE(sio), &s->io);
     sysbus_init_irq(SYS_BUS_DEVICE(sio), &s->irq);
 }
 
diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
index 2c2c7f25b2..84c03dd035 100644
--- a/hw/mips/mips_mipssim.c
+++ b/hw/mips/mips_mipssim.c
@@ -227,7 +227,8 @@ mips_mipssim_init(MachineState *machine)
         qdev_set_legacy_instance_id(dev, 0x3f8, 2);
         qdev_init_nofail(dev);
         sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
-        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->serial.io);
+        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8,
+                      sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0));
     }
 
     if (nd_table[0].used)
-- 
2.24.0

[PATCH v4 21/37] sm501: make SerialMM a child, export chardev property

[Marc-André Lureau]

Embed the SerialMM sybus device, and re-export its "chardev" property.
That way, we can get rid of PROP_PTR "chr-state" and better track
devices relationship.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/display/sm501.c | 33 ++++++++++++++++++++++++---------
 hw/sh4/r2d.c       |  2 +-
 2 files changed, 25 insertions(+), 10 deletions(-)

diff --git a/hw/display/sm501.c b/hw/display/sm501.c
index 1f33c87e65..c4445b28f9 100644
--- a/hw/display/sm501.c
+++ b/hw/display/sm501.c
@@ -1930,13 +1930,14 @@ typedef struct {
     SM501State state;
     uint32_t vram_size;
     uint32_t base;
-    void *chr_state;
+    SerialMM serial;
 } SM501SysBusState;
 
 static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
 {
     SM501SysBusState *s = SYSBUS_SM501(dev);
     SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
+    SerialState *ss = &s->serial.serial;
     DeviceState *usb_dev;
 
     sm501_init(&s->state, dev, s->vram_size);
@@ -1958,17 +1959,19 @@ static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
     sysbus_pass_irq(sbd, SYS_BUS_DEVICE(usb_dev));
 
     /* bridge to serial emulation module */
-    if (s->chr_state) {
-        serial_mm_init(&s->state.mmio_region, SM501_UART0, 2,
-                       NULL, /* TODO : chain irq to IRL */
-                       115200, s->chr_state, DEVICE_LITTLE_ENDIAN);
+    /* FIXME: SM501_UART0 is always mapped, no need to check for the backend */
+    if (qemu_chr_fe_backend_connected(&ss->chr)) {
+        MemoryRegion *mr;
+        qdev_init_nofail(DEVICE(&s->serial));
+        mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->serial), 0);
+        memory_region_add_subregion(&s->state.mmio_region, SM501_UART0, mr);
+        /* TODO : chain irq to IRL */
     }
 }
 
 static Property sm501_sysbus_properties[] = {
     DEFINE_PROP_UINT32("vram-size", SM501SysBusState, vram_size, 0),
     DEFINE_PROP_UINT32("base", SM501SysBusState, base, 0),
-    DEFINE_PROP_PTR("chr-state", SM501SysBusState, chr_state),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -1999,9 +2002,20 @@ static void sm501_sysbus_class_init(ObjectClass *klass, void *data)
     dc->props = sm501_sysbus_properties;
     dc->reset = sm501_reset_sysbus;
     dc->vmsd = &vmstate_sm501_sysbus;
-    /* Note: pointer property "chr-state" may remain null, thus
-     * no need for dc->user_creatable = false;
-     */
+}
+
+static void sm501_sysbus_init(Object *o)
+{
+    SM501SysBusState *sm501 = SYSBUS_SM501(o);
+    SerialMM *smm = &sm501->serial;
+
+    sysbus_init_child_obj(o, "serial", smm, sizeof(SerialMM), TYPE_SERIAL_MM);
+    qdev_set_legacy_instance_id(DEVICE(smm), SM501_UART0, 2);
+    qdev_prop_set_uint8(DEVICE(smm), "regshift", 2);
+    qdev_prop_set_uint8(DEVICE(smm), "endianness", DEVICE_LITTLE_ENDIAN);
+
+    object_property_add_alias(o, "chardev",
+                              OBJECT(smm), "chardev", &error_abort);
 }
 
 static const TypeInfo sm501_sysbus_info = {
@@ -2009,6 +2023,7 @@ static const TypeInfo sm501_sysbus_info = {
     .parent        = TYPE_SYS_BUS_DEVICE,
     .instance_size = sizeof(SM501SysBusState),
     .class_init    = sm501_sysbus_class_init,
+    .instance_init = sm501_sysbus_init,
 };
 
 #define TYPE_PCI_SM501 "sm501"
diff --git a/hw/sh4/r2d.c b/hw/sh4/r2d.c
index ee0840f380..72bb5285cc 100644
--- a/hw/sh4/r2d.c
+++ b/hw/sh4/r2d.c
@@ -272,7 +272,7 @@ static void r2d_init(MachineState *machine)
     busdev = SYS_BUS_DEVICE(dev);
     qdev_prop_set_uint32(dev, "vram-size", SM501_VRAM_SIZE);
     qdev_prop_set_uint32(dev, "base", 0x10000000);
-    qdev_prop_set_ptr(dev, "chr-state", serial_hd(2));
+    qdev_prop_set_chr(dev, "chardev", serial_hd(2));
     qdev_init_nofail(dev);
     sysbus_mmio_map(busdev, 0, 0x10000000);
     sysbus_mmio_map(busdev, 1, 0x13e00000);
-- 
2.24.0

[PATCH v4 22/37] vmmouse: replace PROP_PTR with PROP_LINK

[Marc-André Lureau]

While at it, use the expected type.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/i386/pc.c             | 6 +++---
 hw/i386/vmmouse.c        | 8 +++-----
 hw/input/pckbd.c         | 8 +++-----
 include/hw/input/i8042.h | 4 +++-
 4 files changed, 12 insertions(+), 14 deletions(-)

diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index 49e2259327..d7ac8401a5 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1358,9 +1358,9 @@ static void pc_superio_init(ISABus *isa_bus, bool create_fdctrl, bool no_vmport)
         vmmouse = NULL;
     }
     if (vmmouse) {
-        DeviceState *dev = DEVICE(vmmouse);
-        qdev_prop_set_ptr(dev, "ps2_mouse", i8042);
-        qdev_init_nofail(dev);
+        object_property_set_link(OBJECT(vmmouse), OBJECT(i8042),
+                                 "i8042", &error_abort);
+        qdev_init_nofail(DEVICE(vmmouse));
     }
     port92 = isa_create_simple(isa_bus, "port92");
 
diff --git a/hw/i386/vmmouse.c b/hw/i386/vmmouse.c
index 41ad91ad53..c0c329f817 100644
--- a/hw/i386/vmmouse.c
+++ b/hw/i386/vmmouse.c
@@ -66,7 +66,7 @@ typedef struct VMMouseState
     uint16_t status;
     uint8_t absolute;
     QEMUPutMouseEntry *entry;
-    void *ps2_mouse;
+    ISAKBDState *i8042;
 } VMMouseState;
 
 static uint32_t vmmouse_get_status(VMMouseState *s)
@@ -105,7 +105,7 @@ static void vmmouse_mouse_event(void *opaque, int x, int y, int dz, int buttons_
 
     /* need to still generate PS2 events to notify driver to
        read from queue */
-    i8042_isa_mouse_fake_event(s->ps2_mouse);
+    i8042_isa_mouse_fake_event(s->i8042);
 }
 
 static void vmmouse_remove_handler(VMMouseState *s)
@@ -275,7 +275,7 @@ static void vmmouse_realizefn(DeviceState *dev, Error **errp)
 }
 
 static Property vmmouse_properties[] = {
-    DEFINE_PROP_PTR("ps2_mouse", VMMouseState, ps2_mouse),
+    DEFINE_PROP_LINK("i8042", VMMouseState, i8042, TYPE_I8042, ISAKBDState *),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -287,8 +287,6 @@ static void vmmouse_class_initfn(ObjectClass *klass, void *data)
     dc->reset = vmmouse_reset;
     dc->vmsd = &vmstate_vmmouse;
     dc->props = vmmouse_properties;
-    /* Reason: pointer property "ps2_mouse" */
-    dc->user_creatable = false;
 }
 
 static const TypeInfo vmmouse_info = {
diff --git a/hw/input/pckbd.c b/hw/input/pckbd.c
index f0acfd86f7..9b641021c9 100644
--- a/hw/input/pckbd.c
+++ b/hw/input/pckbd.c
@@ -483,17 +483,15 @@ void i8042_mm_init(qemu_irq kbd_irq, qemu_irq mouse_irq,
 
 #define I8042(obj) OBJECT_CHECK(ISAKBDState, (obj), TYPE_I8042)
 
-typedef struct ISAKBDState {
+struct ISAKBDState {
     ISADevice parent_obj;
 
     KBDState kbd;
     MemoryRegion io[2];
-} ISAKBDState;
+};
 
-void i8042_isa_mouse_fake_event(void *opaque)
+void i8042_isa_mouse_fake_event(ISAKBDState *isa)
 {
-    ISADevice *dev = opaque;
-    ISAKBDState *isa = I8042(dev);
     KBDState *s = &isa->kbd;
 
     ps2_mouse_fake_event(s->mouse);
diff --git a/include/hw/input/i8042.h b/include/hw/input/i8042.h
index 246e6f3335..8eaebf50ce 100644
--- a/include/hw/input/i8042.h
+++ b/include/hw/input/i8042.h
@@ -14,10 +14,12 @@
 
 #define I8042_A20_LINE "a20"
 
+typedef struct ISAKBDState ISAKBDState;
+
 void i8042_mm_init(qemu_irq kbd_irq, qemu_irq mouse_irq,
                    MemoryRegion *region, ram_addr_t size,
                    hwaddr mask);
-void i8042_isa_mouse_fake_event(void *opaque);
+void i8042_isa_mouse_fake_event(ISAKBDState *isa);
 void i8042_setup_a20_line(ISADevice *dev, qemu_irq a20_out);
 
 #endif /* HW_INPUT_I8042_H */
-- 
2.24.0

[PATCH v4 23/37] lance: replace PROP_PTR with PROP_LINK

[Marc-André Lureau]

The device remains non-user creatable since it is a sysbus device.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/dma/sparc32_dma.c | 2 +-
 hw/net/lance.c       | 5 ++---
 hw/net/pcnet-pci.c   | 2 +-
 hw/net/pcnet.h       | 2 +-
 4 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/hw/dma/sparc32_dma.c b/hw/dma/sparc32_dma.c
index 0e5bbcdc7f..3e4da0c47f 100644
--- a/hw/dma/sparc32_dma.c
+++ b/hw/dma/sparc32_dma.c
@@ -346,7 +346,7 @@ static void sparc32_ledma_device_realize(DeviceState *dev, Error **errp)
     d = qdev_create(NULL, TYPE_LANCE);
     object_property_add_child(OBJECT(dev), "lance", OBJECT(d), errp);
     qdev_set_nic_properties(d, nd);
-    qdev_prop_set_ptr(d, "dma", dev);
+    object_property_set_link(OBJECT(d), OBJECT(dev), "dma", errp);
     qdev_init_nofail(d);
 }
 
diff --git a/hw/net/lance.c b/hw/net/lance.c
index 6631e2a4e0..4d96299041 100644
--- a/hw/net/lance.c
+++ b/hw/net/lance.c
@@ -138,7 +138,8 @@ static void lance_instance_init(Object *obj)
 }
 
 static Property lance_properties[] = {
-    DEFINE_PROP_PTR("dma", SysBusPCNetState, state.dma_opaque),
+    DEFINE_PROP_LINK("dma", SysBusPCNetState, state.dma_opaque,
+                     TYPE_DEVICE, DeviceState *),
     DEFINE_NIC_PROPERTIES(SysBusPCNetState, state.conf),
     DEFINE_PROP_END_OF_LIST(),
 };
@@ -153,8 +154,6 @@ static void lance_class_init(ObjectClass *klass, void *data)
     dc->reset = lance_reset;
     dc->vmsd = &vmstate_lance;
     dc->props = lance_properties;
-    /* Reason: pointer property "dma" */
-    dc->user_creatable = false;
 }
 
 static const TypeInfo lance_info = {
diff --git a/hw/net/pcnet-pci.c b/hw/net/pcnet-pci.c
index 4723c30c79..d067d21e2c 100644
--- a/hw/net/pcnet-pci.c
+++ b/hw/net/pcnet-pci.c
@@ -231,7 +231,7 @@ static void pci_pcnet_realize(PCIDevice *pci_dev, Error **errp)
     s->irq = pci_allocate_irq(pci_dev);
     s->phys_mem_read = pci_physical_memory_read;
     s->phys_mem_write = pci_physical_memory_write;
-    s->dma_opaque = pci_dev;
+    s->dma_opaque = DEVICE(pci_dev);
 
     pcnet_common_init(DEVICE(pci_dev), s, &net_pci_pcnet_info);
 }
diff --git a/hw/net/pcnet.h b/hw/net/pcnet.h
index 28d19a5c6f..f49b213c57 100644
--- a/hw/net/pcnet.h
+++ b/hw/net/pcnet.h
@@ -50,7 +50,7 @@ struct PCNetState_st {
                          uint8_t *buf, int len, int do_bswap);
     void (*phys_mem_write)(void *dma_opaque, hwaddr addr,
                           uint8_t *buf, int len, int do_bswap);
-    void *dma_opaque;
+    DeviceState *dma_opaque;
     int tx_busy;
     int looptest;
 };
-- 
2.24.0

[PATCH v4 24/37] etraxfs: remove PROP_PTR usage

[Marc-André Lureau]

etraxfs_dma_client are not Object, so can't be exposed to user with
QOM path. Let's remove property usage and move the constructor to the
.c unit, simplifying some code on the way.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/net/etraxfs_eth.c      | 44 +++++++++++++++++++++++++++++----------
 include/hw/cris/etraxfs.h | 20 +++---------------
 2 files changed, 36 insertions(+), 28 deletions(-)

diff --git a/hw/net/etraxfs_eth.c b/hw/net/etraxfs_eth.c
index 4cfbf1135a..f30d963487 100644
--- a/hw/net/etraxfs_eth.c
+++ b/hw/net/etraxfs_eth.c
@@ -338,14 +338,8 @@ typedef struct ETRAXFSEthState
     uint8_t macaddr[2][6];
     uint32_t regs[FS_ETH_MAX_REGS];
 
-    union {
-        void *vdma_out;
-        struct etraxfs_dma_client *dma_out;
-    };
-    union {
-        void *vdma_in;
-        struct etraxfs_dma_client *dma_in;
-    };
+    struct etraxfs_dma_client *dma_out;
+    struct etraxfs_dma_client *dma_in;
 
     /* MDIO bus.  */
     struct qemu_mdio mdio_bus;
@@ -635,8 +629,6 @@ static void etraxfs_eth_realize(DeviceState *dev, Error **errp)
 
 static Property etraxfs_eth_properties[] = {
     DEFINE_PROP_UINT32("phyaddr", ETRAXFSEthState, phyaddr, 1),
-    DEFINE_PROP_PTR("dma_out", ETRAXFSEthState, vdma_out),
-    DEFINE_PROP_PTR("dma_in", ETRAXFSEthState, vdma_in),
     DEFINE_NIC_PROPERTIES(ETRAXFSEthState, conf),
     DEFINE_PROP_END_OF_LIST(),
 };
@@ -648,10 +640,40 @@ static void etraxfs_eth_class_init(ObjectClass *klass, void *data)
     dc->realize = etraxfs_eth_realize;
     dc->reset = etraxfs_eth_reset;
     dc->props = etraxfs_eth_properties;
-    /* Reason: pointer properties "dma_out", "dma_in" */
+    /* Reason: dma_out, dma_in are not user settable */
     dc->user_creatable = false;
 }
 
+
+/* Instantiate an ETRAXFS Ethernet MAC.  */
+DeviceState *
+etraxfs_eth_init(NICInfo *nd, hwaddr base, int phyaddr,
+                 struct etraxfs_dma_client *dma_out,
+                 struct etraxfs_dma_client *dma_in)
+{
+    DeviceState *dev;
+    qemu_check_nic_model(nd, "fseth");
+
+    dev = qdev_create(NULL, "etraxfs-eth");
+    qdev_set_nic_properties(dev, nd);
+    qdev_prop_set_uint32(dev, "phyaddr", phyaddr);
+
+    /*
+     * TODO: QOM design, define a QOM interface for "I am an etraxfs
+     * DMA client" (which replaces the current 'struct
+     * etraxfs_dma_client' ad-hoc interface), implement it on the
+     * ethernet device, and then have QOM link properties on the DMA
+     * controller device so that you can pass the interface
+     * implementations to it.
+     */
+    ETRAX_FS_ETH(dev)->dma_out = dma_out;
+    ETRAX_FS_ETH(dev)->dma_in = dma_in;
+    qdev_init_nofail(dev);
+    sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, base);
+
+    return dev;
+}
+
 static const TypeInfo etraxfs_eth_info = {
     .name          = TYPE_ETRAX_FS_ETH,
     .parent        = TYPE_SYS_BUS_DEVICE,
diff --git a/include/hw/cris/etraxfs.h b/include/hw/cris/etraxfs.h
index aa146a2cd8..403e7f95e6 100644
--- a/include/hw/cris/etraxfs.h
+++ b/include/hw/cris/etraxfs.h
@@ -30,23 +30,9 @@
 #include "hw/qdev-properties.h"
 #include "hw/sysbus.h"
 
-/* Instantiate an ETRAXFS Ethernet MAC.  */
-static inline DeviceState *
-etraxfs_eth_init(NICInfo *nd, hwaddr base, int phyaddr,
-                 void *dma_out, void *dma_in)
-{
-    DeviceState *dev;
-    qemu_check_nic_model(nd, "fseth");
-
-    dev = qdev_create(NULL, "etraxfs-eth");
-    qdev_set_nic_properties(dev, nd);
-    qdev_prop_set_uint32(dev, "phyaddr", phyaddr);
-    qdev_prop_set_ptr(dev, "dma_out", dma_out);
-    qdev_prop_set_ptr(dev, "dma_in", dma_in);
-    qdev_init_nofail(dev);
-    sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, base);
-    return dev;
-}
+DeviceState *etraxfs_eth_init(NICInfo *nd, hwaddr base, int phyaddr,
+                              struct etraxfs_dma_client *dma_out,
+                              struct etraxfs_dma_client *dma_in);
 
 static inline DeviceState *etraxfs_ser_create(hwaddr addr,
                                               qemu_irq irq,
-- 
2.24.0

[PATCH v4 25/37] dp8393x: replace PROP_PTR with PROP_LINK

[Marc-André Lureau]

Link property is the correct way to pass a MemoryRegion to a device
for DMA purposes.

Sidenote: as a sysbus device, this remains non-usercreatable
even though we can drop the specific flag here.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/m68k/q800.c      | 3 ++-
 hw/mips/mips_jazz.c | 3 ++-
 hw/net/dp8393x.c    | 7 +++----
 3 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/hw/m68k/q800.c b/hw/m68k/q800.c
index 4ca8678007..8f3eb6bfe7 100644
--- a/hw/m68k/q800.c
+++ b/hw/m68k/q800.c
@@ -239,7 +239,8 @@ static void q800_init(MachineState *machine)
     qdev_set_nic_properties(dev, &nd_table[0]);
     qdev_prop_set_uint8(dev, "it_shift", 2);
     qdev_prop_set_bit(dev, "big_endian", true);
-    qdev_prop_set_ptr(dev, "dma_mr", get_system_memory());
+    object_property_set_link(OBJECT(dev), OBJECT(get_system_memory()),
+                             "dma_mr", &error_abort);
     qdev_init_nofail(dev);
     sysbus = SYS_BUS_DEVICE(dev);
     sysbus_mmio_map(sysbus, 0, SONIC_BASE);
diff --git a/hw/mips/mips_jazz.c b/hw/mips/mips_jazz.c
index d978bb64a0..1518eb5e55 100644
--- a/hw/mips/mips_jazz.c
+++ b/hw/mips/mips_jazz.c
@@ -284,7 +284,8 @@ static void mips_jazz_init(MachineState *machine,
             dev = qdev_create(NULL, "dp8393x");
             qdev_set_nic_properties(dev, nd);
             qdev_prop_set_uint8(dev, "it_shift", 2);
-            qdev_prop_set_ptr(dev, "dma_mr", rc4030_dma_mr);
+            object_property_set_link(OBJECT(dev), OBJECT(rc4030_dma_mr),
+                                     "dma_mr", &error_abort);
             qdev_init_nofail(dev);
             sysbus = SYS_BUS_DEVICE(dev);
             sysbus_mmio_map(sysbus, 0, 0x80001000);
diff --git a/hw/net/dp8393x.c b/hw/net/dp8393x.c
index 3d991af163..cdc2631c0c 100644
--- a/hw/net/dp8393x.c
+++ b/hw/net/dp8393x.c
@@ -175,7 +175,7 @@ typedef struct dp8393xState {
     int loopback_packet;
 
     /* Memory access */
-    void *dma_mr;
+    MemoryRegion *dma_mr;
     AddressSpace as;
 } dp8393xState;
 
@@ -948,7 +948,8 @@ static const VMStateDescription vmstate_dp8393x = {
 
 static Property dp8393x_properties[] = {
     DEFINE_NIC_PROPERTIES(dp8393xState, conf),
-    DEFINE_PROP_PTR("dma_mr", dp8393xState, dma_mr),
+    DEFINE_PROP_LINK("dma_mr", dp8393xState, dma_mr,
+                     TYPE_MEMORY_REGION, MemoryRegion *),
     DEFINE_PROP_UINT8("it_shift", dp8393xState, it_shift, 0),
     DEFINE_PROP_BOOL("big_endian", dp8393xState, big_endian, false),
     DEFINE_PROP_END_OF_LIST(),
@@ -963,8 +964,6 @@ static void dp8393x_class_init(ObjectClass *klass, void *data)
     dc->reset = dp8393x_reset;
     dc->vmsd = &vmstate_dp8393x;
     dc->props = dp8393x_properties;
-    /* Reason: dma_mr property can't be set */
-    dc->user_creatable = false;
 }
 
 static const TypeInfo dp8393x_info = {
-- 
2.24.0

[PATCH v4 26/37] leon3: use qemu_irq framework instead of callback as property

[Marc-André Lureau]

"set_pin_in" property is used to define a callback mechanism where the
device says "call the callback function, passing it an opaque cookie
and a 32-bit value". We already have a generic mechanism for doing
that, which is the qemu_irq. So we should just use that.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: KONRAD Frederic <frederic.konrad@adacore.com>
---
 hw/intc/grlib_irqmp.c | 35 ++++-------------------------------
 hw/sparc/leon3.c      | 13 +++++++++----
 target/sparc/cpu.h    |  1 +
 3 files changed, 14 insertions(+), 35 deletions(-)

diff --git a/hw/intc/grlib_irqmp.c b/hw/intc/grlib_irqmp.c
index bc78e1a14f..794c643af2 100644
--- a/hw/intc/grlib_irqmp.c
+++ b/hw/intc/grlib_irqmp.c
@@ -25,6 +25,7 @@
  */
 
 #include "qemu/osdep.h"
+#include "hw/irq.h"
 #include "hw/sysbus.h"
 #include "cpu.h"
 
@@ -58,10 +59,8 @@ typedef struct IRQMP {
 
     MemoryRegion iomem;
 
-    void *set_pil_in;
-    void *set_pil_in_opaque;
-
     IRQMPState *state;
+    qemu_irq irq;
 } IRQMP;
 
 struct IRQMPState {
@@ -82,7 +81,6 @@ static void grlib_irqmp_check_irqs(IRQMPState *state)
     uint32_t      pend   = 0;
     uint32_t      level0 = 0;
     uint32_t      level1 = 0;
-    set_pil_in_fn set_pil_in;
 
     assert(state != NULL);
     assert(state->parent != NULL);
@@ -97,14 +95,8 @@ static void grlib_irqmp_check_irqs(IRQMPState *state)
     trace_grlib_irqmp_check_irqs(state->pending, state->force[0],
                                  state->mask[0], level1, level0);
 
-    set_pil_in = (set_pil_in_fn)state->parent->set_pil_in;
-
     /* Trigger level1 interrupt first and level0 if there is no level1 */
-    if (level1 != 0) {
-        set_pil_in(state->parent->set_pil_in_opaque, level1);
-    } else {
-        set_pil_in(state->parent->set_pil_in_opaque, level0);
-    }
+    qemu_set_irq(state->parent->irq, level1 ?: level0);
 }
 
 static void grlib_irqmp_ack_mask(IRQMPState *state, uint32_t mask)
@@ -335,6 +327,7 @@ static void grlib_irqmp_init(Object *obj)
     IRQMP *irqmp = GRLIB_IRQMP(obj);
     SysBusDevice *dev = SYS_BUS_DEVICE(obj);
 
+    qdev_init_gpio_out_named(DEVICE(obj), &irqmp->irq, "grlib-irq", 1);
     memory_region_init_io(&irqmp->iomem, obj, &grlib_irqmp_ops, irqmp,
                           "irqmp", IRQMP_REG_SIZE);
 
@@ -343,31 +336,11 @@ static void grlib_irqmp_init(Object *obj)
     sysbus_init_mmio(dev, &irqmp->iomem);
 }
 
-static void grlib_irqmp_realize(DeviceState *dev, Error **errp)
-{
-    IRQMP *irqmp = GRLIB_IRQMP(dev);
-
-        /* Check parameters */
-    if (irqmp->set_pil_in == NULL) {
-        error_setg(errp, "set_pil_in cannot be NULL.");
-    }
-}
-
-static Property grlib_irqmp_properties[] = {
-    DEFINE_PROP_PTR("set_pil_in", IRQMP, set_pil_in),
-    DEFINE_PROP_PTR("set_pil_in_opaque", IRQMP, set_pil_in_opaque),
-    DEFINE_PROP_END_OF_LIST(),
-};
-
 static void grlib_irqmp_class_init(ObjectClass *klass, void *data)
 {
     DeviceClass *dc = DEVICE_CLASS(klass);
 
     dc->reset = grlib_irqmp_reset;
-    dc->props = grlib_irqmp_properties;
-    /* Reason: pointer properties "set_pil_in", "set_pil_in_opaque" */
-    dc->user_creatable = false;
-    dc->realize = grlib_irqmp_realize;
 }
 
 static const TypeInfo grlib_irqmp_info = {
diff --git a/hw/sparc/leon3.c b/hw/sparc/leon3.c
index c5f1b1ee72..cac987373e 100644
--- a/hw/sparc/leon3.c
+++ b/hw/sparc/leon3.c
@@ -143,9 +143,14 @@ void leon3_irq_ack(void *irq_manager, int intno)
     grlib_irqmp_ack((DeviceState *)irq_manager, intno);
 }
 
-static void leon3_set_pil_in(void *opaque, uint32_t pil_in)
+/*
+ * This device assumes that the incoming 'level' value on the
+ * qemu_irq is the interrupt number, not just a simple 0/1 level.
+ */
+static void leon3_set_pil_in(void *opaque, int n, int level)
 {
-    CPUSPARCState *env = (CPUSPARCState *)opaque;
+    CPUSPARCState *env = opaque;
+    uint32_t pil_in = level;
     CPUState *cs;
 
     assert(env != NULL);
@@ -225,8 +230,8 @@ static void leon3_generic_hw_init(MachineState *machine)
 
     /* Allocate IRQ manager */
     dev = qdev_create(NULL, TYPE_GRLIB_IRQMP);
-    qdev_prop_set_ptr(dev, "set_pil_in", leon3_set_pil_in);
-    qdev_prop_set_ptr(dev, "set_pil_in_opaque", env);
+    env->pil_irq = qemu_allocate_irq(leon3_set_pil_in, env, 0);
+    qdev_connect_gpio_out_named(dev, "grlib-irq", 0, env->pil_irq);
     qdev_init_nofail(dev);
     sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, LEON3_IRQMP_OFFSET);
     env->irq_manager = dev;
diff --git a/target/sparc/cpu.h b/target/sparc/cpu.h
index ae97c7d9f7..e70fec0133 100644
--- a/target/sparc/cpu.h
+++ b/target/sparc/cpu.h
@@ -541,6 +541,7 @@ struct CPUSPARCState {
 #endif
     sparc_def_t def;
 
+    qemu_irq pil_irq;
     void *irq_manager;
     void (*qemu_irq_ack)(CPUSPARCState *env, void *irq_manager, int intno);
 
-- 
2.24.0

[PATCH v4 27/37] leon3: use qdev gpio facilities for the PIL

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/sparc/leon3.c   | 6 ++++--
 target/sparc/cpu.h | 1 -
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/hw/sparc/leon3.c b/hw/sparc/leon3.c
index cac987373e..1a89d44e57 100644
--- a/hw/sparc/leon3.c
+++ b/hw/sparc/leon3.c
@@ -230,8 +230,10 @@ static void leon3_generic_hw_init(MachineState *machine)
 
     /* Allocate IRQ manager */
     dev = qdev_create(NULL, TYPE_GRLIB_IRQMP);
-    env->pil_irq = qemu_allocate_irq(leon3_set_pil_in, env, 0);
-    qdev_connect_gpio_out_named(dev, "grlib-irq", 0, env->pil_irq);
+    qdev_init_gpio_in_named_with_opaque(DEVICE(env), leon3_set_pil_in,
+                                        env, "pil", 1);
+    qdev_connect_gpio_out_named(dev, "grlib-irq", 0,
+                                qdev_get_gpio_in_named(DEVICE(env), "pil", 0));
     qdev_init_nofail(dev);
     sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, LEON3_IRQMP_OFFSET);
     env->irq_manager = dev;
diff --git a/target/sparc/cpu.h b/target/sparc/cpu.h
index e70fec0133..ae97c7d9f7 100644
--- a/target/sparc/cpu.h
+++ b/target/sparc/cpu.h
@@ -541,7 +541,6 @@ struct CPUSPARCState {
 #endif
     sparc_def_t def;
 
-    qemu_irq pil_irq;
     void *irq_manager;
     void (*qemu_irq_ack)(CPUSPARCState *env, void *irq_manager, int intno);
 
-- 
2.24.0

[PATCH v4 28/37] qdev: use g_strcmp0() instead of open-coding it

[Marc-André Lureau]

Minor code simplification.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/core/qdev.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/hw/core/qdev.c b/hw/core/qdev.c
index cf1ba28fe3..c79befc865 100644
--- a/hw/core/qdev.c
+++ b/hw/core/qdev.c
@@ -394,11 +394,8 @@ static NamedGPIOList *qdev_get_named_gpio_list(DeviceState *dev,
     NamedGPIOList *ngl;
 
     QLIST_FOREACH(ngl, &dev->gpios, node) {
-        /* NULL is a valid and matchable name, otherwise do a normal
-         * strcmp match.
-         */
-        if ((!ngl->name && !name) ||
-                (name && ngl->name && strcmp(name, ngl->name) == 0)) {
+        /* NULL is a valid and matchable name. */
+        if (g_strcmp0(name, ngl->name) == 0) {
             return ngl;
         }
     }
-- 
2.24.0

[PATCH v4 29/37] RFC: mips/cps: fix setting saar property

[Marc-André Lureau]

There is no "saar" property. Note: I haven't been able to test this
code. Help welcome.

May fix commit 043715d1e0fbb3e3411be3f898c5b77b7f90327a ("target/mips:
Update ITU to utilize SAARI and SAAR CP0 registers")

Cc: Aleksandar Markovic <amarkovic@wavecomp.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/mips/cps.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/mips/cps.c b/hw/mips/cps.c
index 1660f86908..c49868d5da 100644
--- a/hw/mips/cps.c
+++ b/hw/mips/cps.c
@@ -106,7 +106,7 @@ static void mips_cps_realize(DeviceState *dev, Error **errp)
         object_property_set_bool(OBJECT(&s->itu), saar_present, "saar-present",
                                  &err);
         if (saar_present) {
-            qdev_prop_set_ptr(DEVICE(&s->itu), "saar", (void *)&env->CP0_SAAR);
+            s->itu.saar = &env->CP0_SAAR;
         }
         object_property_set_bool(OBJECT(&s->itu), true, "realized", &err);
         if (err != NULL) {
-- 
2.24.0

[PATCH v4 30/37] cris: improve passing PIC interrupt vector to the CPU

[Marc-André Lureau]

Instead of accessing cpu interrupt vector directly from PIC, send the
vector value over the qemu_irq.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
---
 hw/cris/axis_dev88.c  |  4 ----
 hw/intc/etraxfs_pic.c | 26 +-------------------------
 target/cris/cpu.c     |  8 ++++++++
 target/cris/cpu.h     |  1 +
 4 files changed, 10 insertions(+), 29 deletions(-)

diff --git a/hw/cris/axis_dev88.c b/hw/cris/axis_dev88.c
index 940c7dd122..be7760476a 100644
--- a/hw/cris/axis_dev88.c
+++ b/hw/cris/axis_dev88.c
@@ -253,7 +253,6 @@ void axisdev88_init(MachineState *machine)
     const char *kernel_filename = machine->kernel_filename;
     const char *kernel_cmdline = machine->kernel_cmdline;
     CRISCPU *cpu;
-    CPUCRISState *env;
     DeviceState *dev;
     SysBusDevice *s;
     DriveInfo *nand;
@@ -267,7 +266,6 @@ void axisdev88_init(MachineState *machine)
 
     /* init CPUs */
     cpu = CRIS_CPU(cpu_create(machine->cpu_type));
-    env = &cpu->env;
 
     /* allocate RAM */
     memory_region_allocate_system_memory(phys_ram, NULL, "axisdev88.ram",
@@ -297,8 +295,6 @@ void axisdev88_init(MachineState *machine)
 
 
     dev = qdev_create(NULL, "etraxfs,pic");
-    /* FIXME: Is there a proper way to signal vectors to the CPU core?  */
-    qdev_prop_set_ptr(dev, "interrupt_vector", &env->interrupt_vector);
     qdev_init_nofail(dev);
     s = SYS_BUS_DEVICE(dev);
     sysbus_mmio_map(s, 0, 0x3001c000);
diff --git a/hw/intc/etraxfs_pic.c b/hw/intc/etraxfs_pic.c
index 77f652acec..12988c7aa9 100644
--- a/hw/intc/etraxfs_pic.c
+++ b/hw/intc/etraxfs_pic.c
@@ -27,8 +27,6 @@
 #include "qemu/module.h"
 #include "hw/irq.h"
 #include "hw/qdev-properties.h"
-//#include "pc.h"
-//#include "etraxfs.h"
 
 #define D(x)
 
@@ -48,7 +46,6 @@ struct etrax_pic
     SysBusDevice parent_obj;
 
     MemoryRegion mmio;
-    void *interrupt_vector;
     qemu_irq parent_irq;
     qemu_irq parent_nmi;
     uint32_t regs[R_MAX];
@@ -79,11 +76,7 @@ static void pic_update(struct etrax_pic *fs)
         }
     }
 
-    if (fs->interrupt_vector) {
-        /* hack alert: ptr property */
-        *(uint32_t*)(fs->interrupt_vector) = vector;
-    }
-    qemu_set_irq(fs->parent_irq, !!vector);
+    qemu_set_irq(fs->parent_irq, vector);
 }
 
 static uint64_t
@@ -163,28 +156,11 @@ static void etraxfs_pic_init(Object *obj)
     sysbus_init_mmio(sbd, &s->mmio);
 }
 
-static Property etraxfs_pic_properties[] = {
-    DEFINE_PROP_PTR("interrupt_vector", struct etrax_pic, interrupt_vector),
-    DEFINE_PROP_END_OF_LIST(),
-};
-
-static void etraxfs_pic_class_init(ObjectClass *klass, void *data)
-{
-    DeviceClass *dc = DEVICE_CLASS(klass);
-
-    dc->props = etraxfs_pic_properties;
-    /*
-     * Note: pointer property "interrupt_vector" may remain null, thus
-     * no need for dc->user_creatable = false;
-     */
-}
-
 static const TypeInfo etraxfs_pic_info = {
     .name          = TYPE_ETRAX_FS_PIC,
     .parent        = TYPE_SYS_BUS_DEVICE,
     .instance_size = sizeof(struct etrax_pic),
     .instance_init = etraxfs_pic_init,
-    .class_init    = etraxfs_pic_class_init,
 };
 
 static void etraxfs_pic_register_types(void)
diff --git a/target/cris/cpu.c b/target/cris/cpu.c
index 7adfd6caf4..6a857f548d 100644
--- a/target/cris/cpu.c
+++ b/target/cris/cpu.c
@@ -147,6 +147,14 @@ static void cris_cpu_set_irq(void *opaque, int irq, int level)
     CPUState *cs = CPU(cpu);
     int type = irq == CRIS_CPU_IRQ ? CPU_INTERRUPT_HARD : CPU_INTERRUPT_NMI;
 
+    if (irq == CRIS_CPU_IRQ) {
+        /*
+         * The PIC passes us the vector for the IRQ as the value it sends
+         * over the qemu_irq line
+         */
+        cpu->env.interrupt_vector = level;
+    }
+
     if (level) {
         cpu_interrupt(cs, type);
     } else {
diff --git a/target/cris/cpu.h b/target/cris/cpu.h
index aba0a66474..a7c2a8e15b 100644
--- a/target/cris/cpu.h
+++ b/target/cris/cpu.h
@@ -34,6 +34,7 @@
 #define CPU_INTERRUPT_NMI       CPU_INTERRUPT_TGT_EXT_3
 
 /* CRUS CPU device objects interrupt lines.  */
+/* PIC passes the vector for the IRQ as the value of it sends over qemu_irq */
 #define CRIS_CPU_IRQ 0
 #define CRIS_CPU_NMI 1
 
-- 
2.24.0

[PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Marc-André Lureau]

Instead, set the initial data field directly.

(the initial data is an array of 256 bytes. As I don't know if it may
change over time, I keep the pointer to original buffer as is, but it
might be worth to consider to copy it instead)

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/i2c/smbus_eeprom.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
index 54c86a0112..533c728b3b 100644
--- a/hw/i2c/smbus_eeprom.c
+++ b/hw/i2c/smbus_eeprom.c
@@ -44,7 +44,7 @@
 typedef struct SMBusEEPROMDevice {
     SMBusDevice smbusdev;
     uint8_t data[SMBUS_EEPROM_SIZE];
-    void *init_data;
+    uint8_t *init_data;
     uint8_t offset;
     bool accessed;
 } SMBusEEPROMDevice;
@@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
 
 static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
 {
+    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
+
     smbus_eeprom_reset(dev);
+    if (eeprom->init_data == NULL) {
+        error_setg(errp, "init_data cannot be NULL");
+    }
 }
 
-static Property smbus_eeprom_properties[] = {
-    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
-    DEFINE_PROP_END_OF_LIST(),
-};
-
 static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
 {
     DeviceClass *dc = DEVICE_CLASS(klass);
@@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
     dc->reset = smbus_eeprom_reset;
     sc->receive_byte = eeprom_receive_byte;
     sc->write_data = eeprom_write_data;
-    dc->props = smbus_eeprom_properties;
     dc->vmsd = &vmstate_smbus_eeprom;
-    /* Reason: pointer property "data" */
+    /* Reason: init_data */
     dc->user_creatable = false;
 }
 
@@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
 
     dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
     qdev_prop_set_uint8(dev, "address", address);
-    qdev_prop_set_ptr(dev, "data", eeprom_buf);
+    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
     qdev_init_nofail(dev);
 }
 
-- 
2.24.0

[PATCH v4 32/37] omap-intc: remove PROP_PTR

[Marc-André Lureau]

Since clocks are not QOM objects, replace PROP_PTR of clocks with
setters methods.

(in theory there should probably be different methods for omap1 &
omap2 intc, but this is left as a future improvement)

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/arm/omap1.c        |  4 ++--
 hw/arm/omap2.c        |  4 ++--
 hw/intc/omap_intc.c   | 17 ++++++++++-------
 include/hw/arm/omap.h | 14 ++++++++++++++
 4 files changed, 28 insertions(+), 11 deletions(-)

diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
index 6ce038a453..1afd1d3d7f 100644
--- a/hw/arm/omap1.c
+++ b/hw/arm/omap1.c
@@ -3889,7 +3889,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
 
     s->ih[0] = qdev_create(NULL, "omap-intc");
     qdev_prop_set_uint32(s->ih[0], "size", 0x100);
-    qdev_prop_set_ptr(s->ih[0], "clk", omap_findclk(s, "arminth_ck"));
+    omap_intc_set_iclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "arminth_ck"));
     qdev_init_nofail(s->ih[0]);
     busdev = SYS_BUS_DEVICE(s->ih[0]);
     sysbus_connect_irq(busdev, 0,
@@ -3899,7 +3899,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
     sysbus_mmio_map(busdev, 0, 0xfffecb00);
     s->ih[1] = qdev_create(NULL, "omap-intc");
     qdev_prop_set_uint32(s->ih[1], "size", 0x800);
-    qdev_prop_set_ptr(s->ih[1], "clk", omap_findclk(s, "arminth_ck"));
+    omap_intc_set_iclk(OMAP_INTC(s->ih[1]), omap_findclk(s, "arminth_ck"));
     qdev_init_nofail(s->ih[1]);
     busdev = SYS_BUS_DEVICE(s->ih[1]);
     sysbus_connect_irq(busdev, 0,
diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
index 457f152bac..1d7cc435ef 100644
--- a/hw/arm/omap2.c
+++ b/hw/arm/omap2.c
@@ -2308,8 +2308,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
     /* Actually mapped at any 2K boundary in the ARM11 private-peripheral if */
     s->ih[0] = qdev_create(NULL, "omap2-intc");
     qdev_prop_set_uint8(s->ih[0], "revision", 0x21);
-    qdev_prop_set_ptr(s->ih[0], "fclk", omap_findclk(s, "mpu_intc_fclk"));
-    qdev_prop_set_ptr(s->ih[0], "iclk", omap_findclk(s, "mpu_intc_iclk"));
+    omap_intc_set_fclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "mpu_intc_fclk"));
+    omap_intc_set_iclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "mpu_intc_iclk"));
     qdev_init_nofail(s->ih[0]);
     busdev = SYS_BUS_DEVICE(s->ih[0]);
     sysbus_connect_irq(busdev, 0,
diff --git a/hw/intc/omap_intc.c b/hw/intc/omap_intc.c
index 854b709ca0..73bb1c2af4 100644
--- a/hw/intc/omap_intc.c
+++ b/hw/intc/omap_intc.c
@@ -38,10 +38,6 @@ struct omap_intr_handler_bank_s {
     unsigned char priority[32];
 };
 
-#define TYPE_OMAP_INTC "common-omap-intc"
-#define OMAP_INTC(obj) \
-    OBJECT_CHECK(struct omap_intr_handler_s, (obj), TYPE_OMAP_INTC)
-
 struct omap_intr_handler_s {
     SysBusDevice parent_obj;
 
@@ -391,9 +387,18 @@ static void omap_intc_realize(DeviceState *dev, Error **errp)
     }
 }
 
+void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk)
+{
+    intc->iclk = clk;
+}
+
+void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk)
+{
+    intc->fclk = clk;
+}
+
 static Property omap_intc_properties[] = {
     DEFINE_PROP_UINT32("size", struct omap_intr_handler_s, size, 0x100),
-    DEFINE_PROP_PTR("clk", struct omap_intr_handler_s, iclk),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -647,8 +652,6 @@ static void omap2_intc_realize(DeviceState *dev, Error **errp)
 static Property omap2_intc_properties[] = {
     DEFINE_PROP_UINT8("revision", struct omap_intr_handler_s,
     revision, 0x21),
-    DEFINE_PROP_PTR("iclk", struct omap_intr_handler_s, iclk),
-    DEFINE_PROP_PTR("fclk", struct omap_intr_handler_s, fclk),
     DEFINE_PROP_END_OF_LIST(),
 };
 
diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
index f3aa670036..bcecf19f89 100644
--- a/include/hw/arm/omap.h
+++ b/include/hw/arm/omap.h
@@ -67,6 +67,20 @@ void omap_clk_setrate(omap_clk clk, int divide, int multiply);
 int64_t omap_clk_getrate(omap_clk clk);
 void omap_clk_reparent(omap_clk clk, omap_clk parent);
 
+/* omap_intc.c */
+#define TYPE_OMAP_INTC "common-omap-intc"
+#define OMAP_INTC(obj)                                              \
+    OBJECT_CHECK(omap_intr_handler, (obj), TYPE_OMAP_INTC)
+
+typedef struct omap_intr_handler_s omap_intr_handler;
+
+/*
+ * TODO: Ideally we should have a clock framework that
+ * let us wire these clocks up with QOM properties or links.
+ */
+void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
+void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
+
 /* OMAP2 l4 Interconnect */
 struct omap_l4_s;
 struct omap_l4_region_s {
-- 
2.24.0

[PATCH v4 33/37] omap-i2c: remove PROP_PTR

[Marc-André Lureau]

Since clocks are not QOM objects, replace PROP_PTR of clocks with
setters methods.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/arm/omap1.c        |  2 +-
 hw/arm/omap2.c        |  8 ++++----
 hw/i2c/omap_i2c.c     | 19 ++++++++++++-------
 include/hw/arm/omap.h | 13 +++++++++++++
 4 files changed, 30 insertions(+), 12 deletions(-)

diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
index 1afd1d3d7f..807e5f70d1 100644
--- a/hw/arm/omap1.c
+++ b/hw/arm/omap1.c
@@ -4030,7 +4030,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
 
     s->i2c[0] = qdev_create(NULL, "omap_i2c");
     qdev_prop_set_uint8(s->i2c[0], "revision", 0x11);
-    qdev_prop_set_ptr(s->i2c[0], "fclk", omap_findclk(s, "mpuper_ck"));
+    omap_i2c_set_fclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "mpuper_ck"));
     qdev_init_nofail(s->i2c[0]);
     busdev = SYS_BUS_DEVICE(s->i2c[0]);
     sysbus_connect_irq(busdev, 0, qdev_get_gpio_in(s->ih[1], OMAP_INT_I2C));
diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
index 1d7cc435ef..171e2d0472 100644
--- a/hw/arm/omap2.c
+++ b/hw/arm/omap2.c
@@ -2425,8 +2425,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
 
     s->i2c[0] = qdev_create(NULL, "omap_i2c");
     qdev_prop_set_uint8(s->i2c[0], "revision", 0x34);
-    qdev_prop_set_ptr(s->i2c[0], "iclk", omap_findclk(s, "i2c1.iclk"));
-    qdev_prop_set_ptr(s->i2c[0], "fclk", omap_findclk(s, "i2c1.fclk"));
+    omap_i2c_set_iclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "i2c1.iclk"));
+    omap_i2c_set_fclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "i2c1.fclk"));
     qdev_init_nofail(s->i2c[0]);
     busdev = SYS_BUS_DEVICE(s->i2c[0]);
     sysbus_connect_irq(busdev, 0,
@@ -2437,8 +2437,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
 
     s->i2c[1] = qdev_create(NULL, "omap_i2c");
     qdev_prop_set_uint8(s->i2c[1], "revision", 0x34);
-    qdev_prop_set_ptr(s->i2c[1], "iclk", omap_findclk(s, "i2c2.iclk"));
-    qdev_prop_set_ptr(s->i2c[1], "fclk", omap_findclk(s, "i2c2.fclk"));
+    omap_i2c_set_iclk(OMAP_I2C(s->i2c[1]), omap_findclk(s, "i2c2.iclk"));
+    omap_i2c_set_fclk(OMAP_I2C(s->i2c[1]), omap_findclk(s, "i2c2.fclk"));
     qdev_init_nofail(s->i2c[1]);
     busdev = SYS_BUS_DEVICE(s->i2c[1]);
     sysbus_connect_irq(busdev, 0,
diff --git a/hw/i2c/omap_i2c.c b/hw/i2c/omap_i2c.c
index 3ba965a58f..3ccbd5cc2c 100644
--- a/hw/i2c/omap_i2c.c
+++ b/hw/i2c/omap_i2c.c
@@ -28,10 +28,7 @@
 #include "qemu/error-report.h"
 #include "qapi/error.h"
 
-#define TYPE_OMAP_I2C "omap_i2c"
-#define OMAP_I2C(obj) OBJECT_CHECK(OMAPI2CState, (obj), TYPE_OMAP_I2C)
-
-typedef struct OMAPI2CState {
+struct OMAPI2CState {
     SysBusDevice parent_obj;
 
     MemoryRegion iomem;
@@ -56,7 +53,7 @@ typedef struct OMAPI2CState {
     uint8_t divider;
     uint8_t times[2];
     uint16_t test;
-} OMAPI2CState;
+};
 
 #define OMAP2_INTR_REV	0x34
 #define OMAP2_GC_REV	0x34
@@ -504,10 +501,18 @@ static void omap_i2c_realize(DeviceState *dev, Error **errp)
     }
 }
 
+void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk)
+{
+    i2c->iclk = clk;
+}
+
+void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk)
+{
+    i2c->fclk = clk;
+}
+
 static Property omap_i2c_properties[] = {
     DEFINE_PROP_UINT8("revision", OMAPI2CState, revision, 0),
-    DEFINE_PROP_PTR("iclk", OMAPI2CState, iclk),
-    DEFINE_PROP_PTR("fclk", OMAPI2CState, fclk),
     DEFINE_PROP_END_OF_LIST(),
 };
 
diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
index bcecf19f89..39a295ba20 100644
--- a/include/hw/arm/omap.h
+++ b/include/hw/arm/omap.h
@@ -81,6 +81,19 @@ typedef struct omap_intr_handler_s omap_intr_handler;
 void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
 void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
 
+/* omap_i2c.c */
+#define TYPE_OMAP_I2C "omap_i2c"
+#define OMAP_I2C(obj) OBJECT_CHECK(OMAPI2CState, (obj), TYPE_OMAP_I2C)
+
+typedef struct OMAPI2CState OMAPI2CState;
+
+/*
+ * TODO: Ideally we should have a clock framework that
+ * let us wire these clocks up with QOM properties or links.
+ */
+void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk);
+void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk);
+
 /* OMAP2 l4 Interconnect */
 struct omap_l4_s;
 struct omap_l4_region_s {
-- 
2.24.0

[PATCH v4 34/37] omap-gpio: remove PROP_PTR

[Marc-André Lureau]

Since clocks are not QOM objects, replace PROP_PTR of clocks with
setters methods.

Move/adapt the existing TODO comment about a clock framework.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/arm/omap1.c        |  2 +-
 hw/arm/omap2.c        | 13 +++++++------
 hw/gpio/omap_gpio.c   | 42 +++++++++++++++---------------------------
 include/hw/arm/omap.h | 33 +++++++++++++++++++++++++++++----
 4 files changed, 52 insertions(+), 38 deletions(-)

diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
index 807e5f70d1..761cc17ea9 100644
--- a/hw/arm/omap1.c
+++ b/hw/arm/omap1.c
@@ -4012,7 +4012,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
 
     s->gpio = qdev_create(NULL, "omap-gpio");
     qdev_prop_set_int32(s->gpio, "mpu_model", s->mpu_model);
-    qdev_prop_set_ptr(s->gpio, "clk", omap_findclk(s, "arm_gpio_ck"));
+    omap_gpio_set_clk(OMAP1_GPIO(s->gpio), omap_findclk(s, "arm_gpio_ck"));
     qdev_init_nofail(s->gpio);
     sysbus_connect_irq(SYS_BUS_DEVICE(s->gpio), 0,
                        qdev_get_gpio_in(s->ih[0], OMAP_INT_GPIO_BANK1));
diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
index 171e2d0472..e1c11de5ce 100644
--- a/hw/arm/omap2.c
+++ b/hw/arm/omap2.c
@@ -2449,13 +2449,14 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
 
     s->gpio = qdev_create(NULL, "omap2-gpio");
     qdev_prop_set_int32(s->gpio, "mpu_model", s->mpu_model);
-    qdev_prop_set_ptr(s->gpio, "iclk", omap_findclk(s, "gpio_iclk"));
-    qdev_prop_set_ptr(s->gpio, "fclk0", omap_findclk(s, "gpio1_dbclk"));
-    qdev_prop_set_ptr(s->gpio, "fclk1", omap_findclk(s, "gpio2_dbclk"));
-    qdev_prop_set_ptr(s->gpio, "fclk2", omap_findclk(s, "gpio3_dbclk"));
-    qdev_prop_set_ptr(s->gpio, "fclk3", omap_findclk(s, "gpio4_dbclk"));
+    omap2_gpio_set_iclk(OMAP2_GPIO(s->gpio), omap_findclk(s, "gpio_iclk"));
+    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 0, omap_findclk(s, "gpio1_dbclk"));
+    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 1, omap_findclk(s, "gpio2_dbclk"));
+    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 2, omap_findclk(s, "gpio3_dbclk"));
+    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 3, omap_findclk(s, "gpio4_dbclk"));
     if (s->mpu_model == omap2430) {
-        qdev_prop_set_ptr(s->gpio, "fclk4", omap_findclk(s, "gpio5_dbclk"));
+        omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 4,
+                            omap_findclk(s, "gpio5_dbclk"));
     }
     qdev_init_nofail(s->gpio);
     busdev = SYS_BUS_DEVICE(s->gpio);
diff --git a/hw/gpio/omap_gpio.c b/hw/gpio/omap_gpio.c
index 41e1aa798c..85c16897ae 100644
--- a/hw/gpio/omap_gpio.c
+++ b/hw/gpio/omap_gpio.c
@@ -40,10 +40,6 @@ struct omap_gpio_s {
     uint16_t pins;
 };
 
-#define TYPE_OMAP1_GPIO "omap-gpio"
-#define OMAP1_GPIO(obj) \
-    OBJECT_CHECK(struct omap_gpif_s, (obj), TYPE_OMAP1_GPIO)
-
 struct omap_gpif_s {
     SysBusDevice parent_obj;
 
@@ -212,10 +208,6 @@ struct omap2_gpio_s {
     uint8_t delay;
 };
 
-#define TYPE_OMAP2_GPIO "omap2-gpio"
-#define OMAP2_GPIO(obj) \
-    OBJECT_CHECK(struct omap2_gpif_s, (obj), TYPE_OMAP2_GPIO)
-
 struct omap2_gpif_s {
     SysBusDevice parent_obj;
 
@@ -747,21 +739,13 @@ static void omap2_gpio_realize(DeviceState *dev, Error **errp)
     }
 }
 
-/* Using qdev pointer properties for the clocks is not ideal.
- * qdev should support a generic means of defining a 'port' with
- * an arbitrary interface for connecting two devices. Then we
- * could reframe the omap clock API in terms of clock ports,
- * and get some type safety. For now the best qdev provides is
- * passing an arbitrary pointer.
- * (It's not possible to pass in the string which is the clock
- * name, because this device does not have the necessary information
- * (ie the struct omap_mpu_state_s*) to do the clockname to pointer
- * translation.)
- */
+void omap_gpio_set_clk(omap_gpif *gpio, omap_clk clk)
+{
+    gpio->clk = clk;
+}
 
 static Property omap_gpio_properties[] = {
     DEFINE_PROP_INT32("mpu_model", struct omap_gpif_s, mpu_model, 0),
-    DEFINE_PROP_PTR("clk", struct omap_gpif_s, clk),
     DEFINE_PROP_END_OF_LIST(),
 };
 
@@ -784,15 +768,19 @@ static const TypeInfo omap_gpio_info = {
     .class_init    = omap_gpio_class_init,
 };
 
+void omap2_gpio_set_iclk(omap2_gpif *gpio, omap_clk clk)
+{
+    gpio->iclk = clk;
+}
+
+void omap2_gpio_set_fclk(omap2_gpif *gpio, uint8_t i, omap_clk clk)
+{
+    assert(i <= 5);
+    gpio->fclk[i] = clk;
+}
+
 static Property omap2_gpio_properties[] = {
     DEFINE_PROP_INT32("mpu_model", struct omap2_gpif_s, mpu_model, 0),
-    DEFINE_PROP_PTR("iclk", struct omap2_gpif_s, iclk),
-    DEFINE_PROP_PTR("fclk0", struct omap2_gpif_s, fclk[0]),
-    DEFINE_PROP_PTR("fclk1", struct omap2_gpif_s, fclk[1]),
-    DEFINE_PROP_PTR("fclk2", struct omap2_gpif_s, fclk[2]),
-    DEFINE_PROP_PTR("fclk3", struct omap2_gpif_s, fclk[3]),
-    DEFINE_PROP_PTR("fclk4", struct omap2_gpif_s, fclk[4]),
-    DEFINE_PROP_PTR("fclk5", struct omap2_gpif_s, fclk[5]),
     DEFINE_PROP_END_OF_LIST(),
 };
 
diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
index 39a295ba20..6be386d0e2 100644
--- a/include/hw/arm/omap.h
+++ b/include/hw/arm/omap.h
@@ -77,6 +77,16 @@ typedef struct omap_intr_handler_s omap_intr_handler;
 /*
  * TODO: Ideally we should have a clock framework that
  * let us wire these clocks up with QOM properties or links.
+ *
+ * qdev should support a generic means of defining a 'port' with
+ * an arbitrary interface for connecting two devices. Then we
+ * could reframe the omap clock API in terms of clock ports,
+ * and get some type safety. For now the best qdev provides is
+ * passing an arbitrary pointer.
+ * (It's not possible to pass in the string which is the clock
+ * name, because this device does not have the necessary information
+ * (ie the struct omap_mpu_state_s*) to do the clockname to pointer
+ * translation.)
  */
 void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
 void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
@@ -87,13 +97,28 @@ void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
 
 typedef struct OMAPI2CState OMAPI2CState;
 
-/*
- * TODO: Ideally we should have a clock framework that
- * let us wire these clocks up with QOM properties or links.
- */
+/* TODO: clock framework (see above) */
 void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk);
 void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk);
 
+/* omap_gpio.c */
+#define TYPE_OMAP1_GPIO "omap-gpio"
+#define OMAP1_GPIO(obj)                                         \
+    OBJECT_CHECK(struct omap_gpif_s, (obj), TYPE_OMAP1_GPIO)
+
+#define TYPE_OMAP2_GPIO "omap2-gpio"
+#define OMAP2_GPIO(obj)                                         \
+    OBJECT_CHECK(struct omap2_gpif_s, (obj), TYPE_OMAP2_GPIO)
+
+typedef struct omap_gpif_s omap_gpif;
+typedef struct omap2_gpif_s omap2_gpif;
+
+/* TODO: clock framework (see above) */
+void omap_gpio_set_clk(omap_gpif *gpio, omap_clk clk);
+
+void omap2_gpio_set_iclk(omap2_gpif *gpio, omap_clk clk);
+void omap2_gpio_set_fclk(omap2_gpif *gpio, uint8_t i, omap_clk clk);
+
 /* OMAP2 l4 Interconnect */
 struct omap_l4_s;
 struct omap_l4_region_s {
-- 
2.24.0

[PATCH v4 35/37] qdev: remove PROP_MEMORY_REGION

[Marc-André Lureau]

PROP_MEMORY_REGION was a derivative of PROP_PTR, added in commit
ed03d749f3f513b8fb0287757cfda2cb6825f063 (qdev: add MemoryRegion
property) and thankfully no longer needed since commit
3eff40dbf44896a8180c86c84dbdefb2eb173fbe (hw/misc: Remove
mmio_interface device).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 include/hw/qdev-properties.h | 2 --
 1 file changed, 2 deletions(-)

diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h
index 690ff07ae2..5bba033b7b 100644
--- a/include/hw/qdev-properties.h
+++ b/include/hw/qdev-properties.h
@@ -213,8 +213,6 @@ extern const PropertyInfo qdev_prop_pcie_link_width;
     DEFINE_PROP_UNSIGNED(_n, _s, _f, 0, qdev_prop_blocksize, uint16_t)
 #define DEFINE_PROP_PCI_HOST_DEVADDR(_n, _s, _f) \
     DEFINE_PROP(_n, _s, _f, qdev_prop_pci_host_devaddr, PCIHostDeviceAddress)
-#define DEFINE_PROP_MEMORY_REGION(_n, _s, _f)             \
-    DEFINE_PROP(_n, _s, _f, qdev_prop_ptr, MemoryRegion *)
 #define DEFINE_PROP_OFF_AUTO_PCIBAR(_n, _s, _f, _d) \
     DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_off_auto_pcibar, \
                         OffAutoPCIBAR)
-- 
2.24.0

[PATCH v4 36/37] qdev: remove QDEV_PROP_PTR

[Marc-André Lureau]

No longer used in the tree. The comment about user_creatable is still
quite relevant, but there is already a similar comment in qdev-core.h.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/core/qdev-properties.c    | 18 ------------------
 include/hw/qdev-properties.h | 22 ----------------------
 2 files changed, 40 deletions(-)

diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c
index be4cb01f0b..6271169d7d 100644
--- a/hw/core/qdev-properties.c
+++ b/hw/core/qdev-properties.c
@@ -469,13 +469,6 @@ const PropertyInfo qdev_prop_string = {
     .set   = set_string,
 };
 
-/* --- pointer --- */
-
-/* Not a proper property, just for dirty hacks.  TODO Remove it!  */
-const PropertyInfo qdev_prop_ptr = {
-    .name  = "ptr",
-};
-
 /* --- mac address --- */
 
 /*
@@ -1133,17 +1126,6 @@ void qdev_prop_set_enum(DeviceState *dev, const char *name, int value)
                             name, &error_abort);
 }
 
-void qdev_prop_set_ptr(DeviceState *dev, const char *name, void *value)
-{
-    Property *prop;
-    void **ptr;
-
-    prop = qdev_prop_find(dev, name);
-    assert(prop && prop->info == &qdev_prop_ptr);
-    ptr = qdev_get_prop_ptr(dev, prop);
-    *ptr = value;
-}
-
 static GPtrArray *global_props(void)
 {
     static GPtrArray *gp;
diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h
index 5bba033b7b..a708c3870b 100644
--- a/include/hw/qdev-properties.h
+++ b/include/hw/qdev-properties.h
@@ -17,7 +17,6 @@ extern const PropertyInfo qdev_prop_size;
 extern const PropertyInfo qdev_prop_string;
 extern const PropertyInfo qdev_prop_chr;
 extern const PropertyInfo qdev_prop_tpm;
-extern const PropertyInfo qdev_prop_ptr;
 extern const PropertyInfo qdev_prop_macaddr;
 extern const PropertyInfo qdev_prop_on_off_auto;
 extern const PropertyInfo qdev_prop_losttickpolicy;
@@ -168,25 +167,6 @@ extern const PropertyInfo qdev_prop_pcie_link_width;
 #define DEFINE_PROP_PCI_DEVFN(_n, _s, _f, _d)                   \
     DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_pci_devfn, int32_t)
 
-/*
- * Please avoid pointer properties.  If you must use them, you must
- * cover them in their device's class init function as follows:
- *
- * - If the property must be set, the device cannot be used with
- *   device_add, so add code like this:
- *   |* Reason: pointer property "NAME-OF-YOUR-PROP" *|
- *   DeviceClass *dc = DEVICE_CLASS(class);
- *   dc->user_creatable = false;
- *
- * - If the property may safely remain null, document it like this:
- *   |*
- *    * Note: pointer property "interrupt_vector" may remain null, thus
- *    * no need for dc->user_creatable = false;
- *    *|
- */
-#define DEFINE_PROP_PTR(_n, _s, _f)             \
-    DEFINE_PROP(_n, _s, _f, qdev_prop_ptr, void*)
-
 #define DEFINE_PROP_CHR(_n, _s, _f)             \
     DEFINE_PROP(_n, _s, _f, qdev_prop_chr, CharBackend)
 #define DEFINE_PROP_STRING(_n, _s, _f)             \
@@ -259,8 +239,6 @@ void qdev_prop_set_drive(DeviceState *dev, const char *name,
 void qdev_prop_set_macaddr(DeviceState *dev, const char *name,
                            const uint8_t *value);
 void qdev_prop_set_enum(DeviceState *dev, const char *name, int value);
-/* FIXME: Remove opaque pointer properties.  */
-void qdev_prop_set_ptr(DeviceState *dev, const char *name, void *value);
 
 void qdev_prop_register_global(GlobalProperty *prop);
 int qdev_prop_check_globals(void);
-- 
2.24.0

[PATCH v4 37/37] qdev/qom: remove some TODO limitations now that PROP_PTR is gone

[Marc-André Lureau]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/core/qdev.c     |  8 --------
 qom/qom-qmp-cmds.c | 10 ----------
 2 files changed, 18 deletions(-)

diff --git a/hw/core/qdev.c b/hw/core/qdev.c
index c79befc865..c7b3f0dec5 100644
--- a/hw/core/qdev.c
+++ b/hw/core/qdev.c
@@ -736,14 +736,6 @@ void qdev_property_add_static(DeviceState *dev, Property *prop,
     if (prop->info->create) {
         prop->info->create(obj, prop, &local_err);
     } else {
-        /*
-         * TODO qdev_prop_ptr does not have getters or setters.  It must
-         * go now that it can be replaced with links.  The test should be
-         * removed along with it: all static properties are read/write.
-         */
-        if (!prop->info->get && !prop->info->set) {
-            return;
-        }
         object_property_add(obj, prop->name, prop->info->name,
                             prop->info->get, prop->info->set,
                             prop->info->release,
diff --git a/qom/qom-qmp-cmds.c b/qom/qom-qmp-cmds.c
index 236e8e41dd..f4494f98ac 100644
--- a/qom/qom-qmp-cmds.c
+++ b/qom/qom-qmp-cmds.c
@@ -142,16 +142,6 @@ static ObjectPropertyInfo *make_device_property_info(ObjectClass *klass,
                 continue;
             }
 
-            /*
-             * TODO Properties without a parser are just for dirty hacks.
-             * qdev_prop_ptr is the only such PropertyInfo.  It's marked
-             * for removal.  This conditional should be removed along with
-             * it.
-             */
-            if (!prop->info->set && !prop->info->create) {
-                return NULL;           /* no way to set it, don't show */
-            }
-
             info = g_malloc0(sizeof(*info));
             info->name = g_strdup(prop->name);
             info->type = default_type ? g_strdup(default_type)
-- 
2.24.0

Re: [PATCH v4 10/37] serial: realize the serial device

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Instead of calling serial_realize_core(), use the QDev realize
> callback.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/char/serial-isa.c       | 2 +-
>   hw/char/serial-pci-multi.c | 2 +-
>   hw/char/serial-pci.c       | 2 +-
>   hw/char/serial.c           | 8 ++++----
>   include/hw/char/serial.h   | 1 -
>   5 files changed, 7 insertions(+), 8 deletions(-)
> 
> diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
> index 2a4c8de1bf..db8644551e 100644
> --- a/hw/char/serial-isa.c
> +++ b/hw/char/serial-isa.c
> @@ -74,7 +74,7 @@ static void serial_isa_realizefn(DeviceState *dev, Error **errp)
>       index++;
>   
>       isa_init_irq(isadev, &s->irq, isa->isairq);
> -    serial_realize_core(s, errp);
> +    object_property_set_bool(OBJECT(s), true, "realized", errp);
>       qdev_set_legacy_instance_id(dev, isa->iobase, 3);
>   
>       memory_region_init_io(&s->io, OBJECT(isa), &serial_io_ops, s, "serial", 8);
> diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
> index 4891f32230..0e1fdb75d2 100644
> --- a/hw/char/serial-pci-multi.c
> +++ b/hw/char/serial-pci-multi.c
> @@ -106,7 +106,7 @@ static void multi_serial_pci_realize(PCIDevice *dev, Error **errp)
>   
>       for (i = 0; i < nports; i++) {
>           s = pci->state + i;
> -        serial_realize_core(s, &err);
> +        object_property_set_bool(OBJECT(s), true, "realized", &err);
>           if (err != NULL) {
>               error_propagate(errp, err);
>               multi_serial_pci_exit(dev);
> diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
> index db2c17aafd..4b6a217365 100644
> --- a/hw/char/serial-pci.c
> +++ b/hw/char/serial-pci.c
> @@ -49,7 +49,7 @@ static void serial_pci_realize(PCIDevice *dev, Error **errp)
>       SerialState *s = &pci->state;
>       Error *err = NULL;
>   
> -    serial_realize_core(s, &err);
> +    object_property_set_bool(OBJECT(s), true, "realized", &err);
>       if (err != NULL) {
>           error_propagate(errp, err);
>           return;
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 4a598ead21..1f3859eef1 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -934,8 +934,10 @@ static int serial_be_change(void *opaque)
>       return 0;
>   }
>   
> -void serial_realize_core(SerialState *s, Error **errp)
> +static void serial_realize(DeviceState *dev, Error **errp)
>   {
> +    SerialState *s = SERIAL(dev);
> +
>       s->modem_status_poll = timer_new_ns(QEMU_CLOCK_VIRTUAL, (QEMUTimerCB *) serial_update_msl, s);
>   
>       s->fifo_timeout_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, (QEMUTimerCB *) fifo_timeout_int, s);
> @@ -990,7 +992,6 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
>       s->irq = irq;
>       qdev_prop_set_uint32(dev, "baudbase", baudbase);
>       qdev_prop_set_chr(dev, "chardev", chr);
> -    serial_realize_core(s, &error_fatal);
>       qdev_set_legacy_instance_id(dev, base, 2);
>       qdev_init_nofail(dev);
>   
> @@ -1011,6 +1012,7 @@ static void serial_class_init(ObjectClass *klass, void* data)
>       DeviceClass *dc = DEVICE_CLASS(klass);
>   
>       dc->user_creatable = false;
> +    dc->realize = serial_realize;
>       dc->vmsd = &vmstate_serial;
>       dc->props = serial_properties;
>   }
> @@ -1074,8 +1076,6 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
>       s->irq = irq;
>       qdev_prop_set_uint32(dev, "baudbase", baudbase);
>       qdev_prop_set_chr(dev, "chardev", chr);
> -
> -    serial_realize_core(s, &error_fatal);
>       qdev_set_legacy_instance_id(dev, base, 2);
>       qdev_init_nofail(dev);
>   
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index 3dc618598e..571aab97c8 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -83,7 +83,6 @@ typedef struct SerialState {
>   extern const VMStateDescription vmstate_serial;
>   extern const MemoryRegionOps serial_io_ops;
>   
> -void serial_realize_core(SerialState *s, Error **errp);
>   void serial_exit_core(SerialState *s);
>   void serial_set_frequency(SerialState *s, uint32_t frequency);
>   
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 17/37] mips: inline serial_init()

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> The function is specific to mipssim, let's inline it.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/char/serial.c         | 16 ----------------
>   hw/mips/mips_mipssim.c   | 15 ++++++++++++---
>   include/hw/char/serial.h |  2 --
>   3 files changed, 12 insertions(+), 21 deletions(-)
> 
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 164146ede8..23f0b02516 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
>       .class_init = serial_io_class_init,
>   };
>   
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                         Chardev *chr, MemoryRegion *system_io)
> -{
> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> -
> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> -    qdev_init_nofail(DEVICE(sio));
> -
> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
> -
> -    return sio;
> -}
> -
>   static Property serial_properties[] = {
>       DEFINE_PROP_CHR("chardev", SerialState, chr),
>       DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 282bbecb24..bfafa4d7e9 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -40,6 +40,7 @@
>   #include "hw/loader.h"
>   #include "elf.h"
>   #include "hw/sysbus.h"
> +#include "hw/qdev-properties.h"
>   #include "exec/address-spaces.h"
>   #include "qemu/error-report.h"
>   #include "sysemu/qtest.h"
> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
>        * A single 16450 sits at offset 0x3f8. It is attached to
>        * MIPS CPU INT2, which is interrupt 4.
>        */
> -    if (serial_hd(0))
> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
> -                    get_system_io());
> +    if (serial_hd(0)) {
> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> +
> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> +        qdev_init_nofail(dev);
> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> +        memory_region_add_subregion(get_system_io(), 0x3f8,
> +                                    &SERIAL_IO(dev)->serial.io);
> +    }
>   
>       if (nd_table[0].used)
>           /* MIPSnet uses the MIPS CPU INT0, which is interrupt 2. */
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index d356ba838c..535fa23a2b 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -108,8 +108,6 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
>   #define TYPE_SERIAL_IO "serial-io"
>   #define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
>   
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                      Chardev *chr, MemoryRegion *system_io);
>   SerialMM *serial_mm_init(MemoryRegion *address_space,
>                            hwaddr base, int regshift,
>                            qemu_irq irq, int baudbase,
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 16/37] serial: make SerialIO a sysbus device

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Make serial IO a proper sysbus device, similar to serial MM.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/char/serial.c         | 59 ++++++++++++++++++++++++++++++++--------
>   include/hw/char/serial.h | 13 +++++++--
>   2 files changed, 59 insertions(+), 13 deletions(-)
> 
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index e0f1a99e8b..164146ede8 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -985,22 +985,58 @@ const MemoryRegionOps serial_io_ops = {
>       .endianness = DEVICE_LITTLE_ENDIAN,
>   };
>   
> -SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> -                         Chardev *chr, MemoryRegion *system_io)
> +static void serial_io_realize(DeviceState *dev, Error **errp)
>   {
> -    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
> -    SerialState *s = SERIAL(dev);
> +    SerialIO *sio = SERIAL_IO(dev);
> +    SerialState *s = &sio->serial;
>   
> -    s->irq = irq;
> -    qdev_prop_set_uint32(dev, "baudbase", baudbase);
> -    qdev_prop_set_chr(dev, "chardev", chr);
> -    qdev_set_legacy_instance_id(dev, base, 2);
> -    qdev_init_nofail(dev);
> +    qdev_init_nofail(DEVICE(s));
>   
>       memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
> -    memory_region_add_subregion(system_io, base, &s->io);
> +    sysbus_init_irq(SYS_BUS_DEVICE(sio), &s->irq);
> +}
> +
> +static void serial_io_class_init(ObjectClass *klass, void* data)
> +{
> +    DeviceClass *dc = DEVICE_CLASS(klass);
> +
> +    dc->realize = serial_io_realize;
> +    /* No dc->vmsd: class has no migratable state */
> +}
> +
> +static void serial_io_instance_init(Object *o)
> +{
> +    SerialIO *sio = SERIAL_IO(o);
> +
> +    object_initialize_child(o, "serial", &sio->serial, sizeof(sio->serial),
> +                            TYPE_SERIAL, &error_abort, NULL);
> +
> +    qdev_alias_all_properties(DEVICE(&sio->serial), o);
> +}
> +
> +
> +static const TypeInfo serial_io_info = {
> +    .name = TYPE_SERIAL_IO,
> +    .parent = TYPE_SYS_BUS_DEVICE,
> +    .instance_size = sizeof(SerialIO),
> +    .instance_init = serial_io_instance_init,
> +    .class_init = serial_io_class_init,
> +};
> +
> +SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> +                         Chardev *chr, MemoryRegion *system_io)
> +{
> +    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> +
> +    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> +    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> +    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> +    qdev_init_nofail(DEVICE(sio));
> +
> +    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> +    memory_region_add_subregion(system_io, base, &sio->serial.io);
>   
> -    return s;
> +    return sio;
>   }
>   
>   static Property serial_properties[] = {
> @@ -1142,6 +1178,7 @@ static const TypeInfo serial_mm_info = {
>   static void serial_register_types(void)
>   {
>       type_register_static(&serial_info);
> +    type_register_static(&serial_io_info);
>       type_register_static(&serial_mm_info);
>   }
>   
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index d586924937..d356ba838c 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -88,6 +88,12 @@ typedef struct SerialMM {
>       uint8_t endianness;
>   } SerialMM;
>   
> +typedef struct SerialIO {
> +    SysBusDevice parent;
> +
> +    SerialState serial;
> +} SerialIO;
> +
>   extern const VMStateDescription vmstate_serial;
>   extern const MemoryRegionOps serial_io_ops;
>   
> @@ -99,8 +105,11 @@ void serial_set_frequency(SerialState *s, uint32_t frequency);
>   #define TYPE_SERIAL_MM "serial-mm"
>   #define SERIAL_MM(s) OBJECT_CHECK(SerialMM, (s), TYPE_SERIAL_MM)
>   
> -SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> -                         Chardev *chr, MemoryRegion *system_io);
> +#define TYPE_SERIAL_IO "serial-io"
> +#define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
> +
> +SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> +                      Chardev *chr, MemoryRegion *system_io);
>   SerialMM *serial_mm_init(MemoryRegion *address_space,
>                            hwaddr base, int regshift,
>                            qemu_irq irq, int baudbase,
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 28/37] qdev: use g_strcmp0() instead of open-coding it

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Minor code simplification.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/core/qdev.c | 7 ++-----
>   1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index cf1ba28fe3..c79befc865 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -394,11 +394,8 @@ static NamedGPIOList *qdev_get_named_gpio_list(DeviceState *dev,
>       NamedGPIOList *ngl;
>   
>       QLIST_FOREACH(ngl, &dev->gpios, node) {
> -        /* NULL is a valid and matchable name, otherwise do a normal
> -         * strcmp match.
> -         */
> -        if ((!ngl->name && !name) ||
> -                (name && ngl->name && strcmp(name, ngl->name) == 0)) {
> +        /* NULL is a valid and matchable name. */
> +        if (g_strcmp0(name, ngl->name) == 0) {

Cocci pattern?

>               return ngl;
>           }
>       }
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 30/37] cris: improve passing PIC interrupt vector to the CPU

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Instead of accessing cpu interrupt vector directly from PIC, send the
> vector value over the qemu_irq.
> 
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>   hw/cris/axis_dev88.c  |  4 ----
>   hw/intc/etraxfs_pic.c | 26 +-------------------------
>   target/cris/cpu.c     |  8 ++++++++
>   target/cris/cpu.h     |  1 +
>   4 files changed, 10 insertions(+), 29 deletions(-)
> 
> diff --git a/hw/cris/axis_dev88.c b/hw/cris/axis_dev88.c
> index 940c7dd122..be7760476a 100644
> --- a/hw/cris/axis_dev88.c
> +++ b/hw/cris/axis_dev88.c
> @@ -253,7 +253,6 @@ void axisdev88_init(MachineState *machine)
>       const char *kernel_filename = machine->kernel_filename;
>       const char *kernel_cmdline = machine->kernel_cmdline;
>       CRISCPU *cpu;
> -    CPUCRISState *env;
>       DeviceState *dev;
>       SysBusDevice *s;
>       DriveInfo *nand;
> @@ -267,7 +266,6 @@ void axisdev88_init(MachineState *machine)
>   
>       /* init CPUs */
>       cpu = CRIS_CPU(cpu_create(machine->cpu_type));
> -    env = &cpu->env;
>   
>       /* allocate RAM */
>       memory_region_allocate_system_memory(phys_ram, NULL, "axisdev88.ram",
> @@ -297,8 +295,6 @@ void axisdev88_init(MachineState *machine)
>   
>   
>       dev = qdev_create(NULL, "etraxfs,pic");
> -    /* FIXME: Is there a proper way to signal vectors to the CPU core?  */
> -    qdev_prop_set_ptr(dev, "interrupt_vector", &env->interrupt_vector);
>       qdev_init_nofail(dev);
>       s = SYS_BUS_DEVICE(dev);
>       sysbus_mmio_map(s, 0, 0x3001c000);
> diff --git a/hw/intc/etraxfs_pic.c b/hw/intc/etraxfs_pic.c
> index 77f652acec..12988c7aa9 100644
> --- a/hw/intc/etraxfs_pic.c
> +++ b/hw/intc/etraxfs_pic.c
> @@ -27,8 +27,6 @@
>   #include "qemu/module.h"
>   #include "hw/irq.h"
>   #include "hw/qdev-properties.h"
> -//#include "pc.h"
> -//#include "etraxfs.h"
>   
>   #define D(x)
>   
> @@ -48,7 +46,6 @@ struct etrax_pic
>       SysBusDevice parent_obj;
>   
>       MemoryRegion mmio;
> -    void *interrupt_vector;
>       qemu_irq parent_irq;
>       qemu_irq parent_nmi;
>       uint32_t regs[R_MAX];
> @@ -79,11 +76,7 @@ static void pic_update(struct etrax_pic *fs)
>           }
>       }
>   
> -    if (fs->interrupt_vector) {
> -        /* hack alert: ptr property */
> -        *(uint32_t*)(fs->interrupt_vector) = vector;
> -    }
> -    qemu_set_irq(fs->parent_irq, !!vector);
> +    qemu_set_irq(fs->parent_irq, vector);
>   }
>   
>   static uint64_t
> @@ -163,28 +156,11 @@ static void etraxfs_pic_init(Object *obj)
>       sysbus_init_mmio(sbd, &s->mmio);
>   }
>   
> -static Property etraxfs_pic_properties[] = {
> -    DEFINE_PROP_PTR("interrupt_vector", struct etrax_pic, interrupt_vector),
> -    DEFINE_PROP_END_OF_LIST(),
> -};
> -
> -static void etraxfs_pic_class_init(ObjectClass *klass, void *data)
> -{
> -    DeviceClass *dc = DEVICE_CLASS(klass);
> -
> -    dc->props = etraxfs_pic_properties;
> -    /*
> -     * Note: pointer property "interrupt_vector" may remain null, thus
> -     * no need for dc->user_creatable = false;
> -     */
> -}
> -
>   static const TypeInfo etraxfs_pic_info = {
>       .name          = TYPE_ETRAX_FS_PIC,
>       .parent        = TYPE_SYS_BUS_DEVICE,
>       .instance_size = sizeof(struct etrax_pic),
>       .instance_init = etraxfs_pic_init,
> -    .class_init    = etraxfs_pic_class_init,
>   };
>   
>   static void etraxfs_pic_register_types(void)
> diff --git a/target/cris/cpu.c b/target/cris/cpu.c
> index 7adfd6caf4..6a857f548d 100644
> --- a/target/cris/cpu.c
> +++ b/target/cris/cpu.c
> @@ -147,6 +147,14 @@ static void cris_cpu_set_irq(void *opaque, int irq, int level)
>       CPUState *cs = CPU(cpu);
>       int type = irq == CRIS_CPU_IRQ ? CPU_INTERRUPT_HARD : CPU_INTERRUPT_NMI;
>   
> +    if (irq == CRIS_CPU_IRQ) {
> +        /*
> +         * The PIC passes us the vector for the IRQ as the value it sends
> +         * over the qemu_irq line
> +         */
> +        cpu->env.interrupt_vector = level;

I worked on a series that change level to a boolean, but having a way to 
pass a vectored IRQ via the IRQ API seems useful.

Maybe we should clarify the qemu_irq_handler prototype documentation. 
Unfortunately it is declared in "qemu/typedefs.h", and the documentation 
is expected in "hw/irq.h".

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> +    }
> +
>       if (level) {
>           cpu_interrupt(cs, type);
>       } else {
> diff --git a/target/cris/cpu.h b/target/cris/cpu.h
> index aba0a66474..a7c2a8e15b 100644
> --- a/target/cris/cpu.h
> +++ b/target/cris/cpu.h
> @@ -34,6 +34,7 @@
>   #define CPU_INTERRUPT_NMI       CPU_INTERRUPT_TGT_EXT_3
>   
>   /* CRUS CPU device objects interrupt lines.  */
> +/* PIC passes the vector for the IRQ as the value of it sends over qemu_irq */
>   #define CRIS_CPU_IRQ 0
>   #define CRIS_CPU_NMI 1
>   
> 

Re: [PATCH v4 06/37] serial: initial qom-ification

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Make SerialState a device (the following patches will introduce IO/MM
> sysbus serial devices)
> 
> None of the serial_{,mm}_init() callers actually free the returned
> value (even if they did, it would be quite harmless), so we can change
> the object allocation at will.
> 
> However, the devices that embed SerialState must now have their field
> QOM-initialized manually (isa, pci, pci-multi).
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/char/serial-isa.c       |  9 +++++++++
>   hw/char/serial-pci-multi.c | 15 +++++++++++++++
>   hw/char/serial-pci.c       | 13 ++++++++++++-
>   hw/char/serial.c           | 33 +++++++++++++++++++++++++++------
>   include/hw/char/serial.h   |  7 ++++++-
>   5 files changed, 69 insertions(+), 8 deletions(-)
> 
> diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
> index 9e31c51bb6..9a5928b3ee 100644
> --- a/hw/char/serial-isa.c
> +++ b/hw/char/serial-isa.c
> @@ -111,10 +111,19 @@ static void serial_isa_class_initfn(ObjectClass *klass, void *data)
>       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
>   }
>   
> +static void serial_isa_initfn(Object *o)
> +{
> +    ISASerialState *self = ISA_SERIAL(o);
> +
> +    object_initialize_child(o, "serial", &self->state, sizeof(self->state),
> +                            TYPE_SERIAL, &error_abort, NULL);
> +}
> +
>   static const TypeInfo serial_isa_info = {
>       .name          = TYPE_ISA_SERIAL,
>       .parent        = TYPE_ISA_DEVICE,
>       .instance_size = sizeof(ISASerialState),
> +    .instance_init = serial_isa_initfn,
>       .class_init    = serial_isa_class_initfn,
>   };
>   
> diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
> index 5c553c30d0..edfbfdca9e 100644
> --- a/hw/char/serial-pci-multi.c
> +++ b/hw/char/serial-pci-multi.c
> @@ -181,10 +181,24 @@ static void multi_4x_serial_pci_class_initfn(ObjectClass *klass, void *data)
>       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
>   }
>   
> +static void multi_serial_init(Object *o)
> +{
> +    PCIDevice *dev = PCI_DEVICE(o);
> +    PCIMultiSerialState *pms = DO_UPCAST(PCIMultiSerialState, dev, dev);
> +    size_t i, nports = multi_serial_get_port_count(PCI_DEVICE_GET_CLASS(dev));
> +
> +    for (i = 0; i < nports; i++) {
> +        object_initialize_child(o, "serial[*]", &pms->state[i],
> +                                sizeof(pms->state[i]),
> +                                TYPE_SERIAL, &error_abort, NULL);
> +    }
> +}
> +
>   static const TypeInfo multi_2x_serial_pci_info = {
>       .name          = "pci-serial-2x",
>       .parent        = TYPE_PCI_DEVICE,
>       .instance_size = sizeof(PCIMultiSerialState),
> +    .instance_init = multi_serial_init,
>       .class_init    = multi_2x_serial_pci_class_initfn,
>       .interfaces = (InterfaceInfo[]) {
>           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> @@ -196,6 +210,7 @@ static const TypeInfo multi_4x_serial_pci_info = {
>       .name          = "pci-serial-4x",
>       .parent        = TYPE_PCI_DEVICE,
>       .instance_size = sizeof(PCIMultiSerialState),
> +    .instance_init = multi_serial_init,
>       .class_init    = multi_4x_serial_pci_class_initfn,
>       .interfaces = (InterfaceInfo[]) {
>           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
> index cb9b76e22b..f99b6c19e0 100644
> --- a/hw/char/serial-pci.c
> +++ b/hw/char/serial-pci.c
> @@ -40,6 +40,8 @@ typedef struct PCISerialState {
>       uint8_t prog_if;
>   } PCISerialState;
>   
> +#define TYPE_PCI_SERIAL "pci-serial"
> +#define PCI_SERIAL(s) OBJECT_CHECK(PCISerialState, (s), TYPE_PCI_SERIAL)
>   
>   static void serial_pci_realize(PCIDevice *dev, Error **errp)
>   {
> @@ -103,10 +105,19 @@ static void serial_pci_class_initfn(ObjectClass *klass, void *data)
>       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
>   }
>   
> +static void serial_pci_init(Object *o)
> +{
> +    PCISerialState *ps = PCI_SERIAL(o);
> +
> +    object_initialize_child(o, "serial", &ps->state, sizeof(ps->state),
> +                            TYPE_SERIAL, &error_abort, NULL);
> +}
> +
>   static const TypeInfo serial_pci_info = {
> -    .name          = "pci-serial",
> +    .name          = TYPE_PCI_SERIAL,
>       .parent        = TYPE_PCI_DEVICE,
>       .instance_size = sizeof(PCISerialState),
> +    .instance_init = serial_pci_init,
>       .class_init    = serial_pci_class_initfn,
>       .interfaces = (InterfaceInfo[]) {
>           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index b4aa250950..e0a5bec290 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -983,9 +983,8 @@ const MemoryRegionOps serial_io_ops = {
>   SerialState *serial_init(int base, qemu_irq irq, int baudbase,
>                            Chardev *chr, MemoryRegion *system_io)
>   {
> -    SerialState *s;
> -
> -    s = g_malloc0(sizeof(SerialState));
> +    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
> +    SerialState *s = SERIAL(dev);
>   
>       s->irq = irq;
>       s->baudbase = baudbase;
> @@ -993,6 +992,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
>       serial_realize_core(s, &error_fatal);
>   
>       vmstate_register(NULL, base, &vmstate_serial, s);
> +    qdev_init_nofail(dev);
>   
>       memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
>       memory_region_add_subregion(system_io, base, &s->io);
> @@ -1000,6 +1000,20 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
>       return s;
>   }
>   
> +static void serial_class_init(ObjectClass *klass, void *data)
> +{
> +    DeviceClass *dc = DEVICE_CLASS(klass);
> +
> +    dc->user_creatable = false;

Why? Due to the "chardev" property?

> +}
> +
> +static const TypeInfo serial_info = {
> +    .name = TYPE_SERIAL,
> +    .parent = TYPE_DEVICE,
> +    .instance_size = sizeof(SerialState),
> +    .class_init = serial_class_init,
> +};
> +
>   /* Memory mapped interface */
>   static uint64_t serial_mm_read(void *opaque, hwaddr addr,
>                                  unsigned size)
> @@ -1045,9 +1059,8 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
>                               qemu_irq irq, int baudbase,
>                               Chardev *chr, enum device_endian end)
>   {
> -    SerialState *s;
> -
> -    s = g_malloc0(sizeof(SerialState));
> +    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
> +    SerialState *s = SERIAL(dev);
>   
>       s->it_shift = it_shift;
>       s->irq = irq;
> @@ -1056,9 +1069,17 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
>   
>       serial_realize_core(s, &error_fatal);
>       vmstate_register(NULL, base, &vmstate_serial, s);
> +    qdev_init_nofail(dev);
>   
>       memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], s,
>                             "serial", 8 << it_shift);
>       memory_region_add_subregion(address_space, base, &s->io);
>       return s;
>   }
> +
> +static void serial_register_types(void)
> +{
> +    type_register_static(&serial_info);
> +}
> +
> +type_init(serial_register_types)
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index 8be3d8a4f9..180cc7c24e 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -30,10 +30,13 @@
>   #include "exec/memory.h"
>   #include "qemu/fifo8.h"
>   #include "chardev/char.h"
> +#include "hw/sysbus.h"
>   
>   #define UART_FIFO_LENGTH    16      /* 16550A Fifo Length */
>   
>   typedef struct SerialState {
> +    DeviceState parent;
> +
>       uint16_t divider;
>       uint8_t rbr; /* receive register */
>       uint8_t thr; /* transmit holding register */
> @@ -84,7 +87,9 @@ void serial_realize_core(SerialState *s, Error **errp);
>   void serial_exit_core(SerialState *s);
>   void serial_set_frequency(SerialState *s, uint32_t frequency);
>   
> -/* legacy pre qom */
> +#define TYPE_SERIAL "serial"
> +#define SERIAL(s) OBJECT_CHECK(SerialState, (s), TYPE_SERIAL)
> +
>   SerialState *serial_init(int base, qemu_irq irq, int baudbase,
>                            Chardev *chr, MemoryRegion *system_io);
>   SerialState *serial_mm_init(MemoryRegion *address_space,
> 

Re: [PATCH v4 06/37] serial: initial qom-ification

[Marc-André Lureau]

Hi

On Wed, Nov 20, 2019 at 8:42 PM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> > Make SerialState a device (the following patches will introduce IO/MM
> > sysbus serial devices)
> >
> > None of the serial_{,mm}_init() callers actually free the returned
> > value (even if they did, it would be quite harmless), so we can change
> > the object allocation at will.
> >
> > However, the devices that embed SerialState must now have their field
> > QOM-initialized manually (isa, pci, pci-multi).
> >
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >   hw/char/serial-isa.c       |  9 +++++++++
> >   hw/char/serial-pci-multi.c | 15 +++++++++++++++
> >   hw/char/serial-pci.c       | 13 ++++++++++++-
> >   hw/char/serial.c           | 33 +++++++++++++++++++++++++++------
> >   include/hw/char/serial.h   |  7 ++++++-
> >   5 files changed, 69 insertions(+), 8 deletions(-)
> >
> > diff --git a/hw/char/serial-isa.c b/hw/char/serial-isa.c
> > index 9e31c51bb6..9a5928b3ee 100644
> > --- a/hw/char/serial-isa.c
> > +++ b/hw/char/serial-isa.c
> > @@ -111,10 +111,19 @@ static void serial_isa_class_initfn(ObjectClass *klass, void *data)
> >       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
> >   }
> >
> > +static void serial_isa_initfn(Object *o)
> > +{
> > +    ISASerialState *self = ISA_SERIAL(o);
> > +
> > +    object_initialize_child(o, "serial", &self->state, sizeof(self->state),
> > +                            TYPE_SERIAL, &error_abort, NULL);
> > +}
> > +
> >   static const TypeInfo serial_isa_info = {
> >       .name          = TYPE_ISA_SERIAL,
> >       .parent        = TYPE_ISA_DEVICE,
> >       .instance_size = sizeof(ISASerialState),
> > +    .instance_init = serial_isa_initfn,
> >       .class_init    = serial_isa_class_initfn,
> >   };
> >
> > diff --git a/hw/char/serial-pci-multi.c b/hw/char/serial-pci-multi.c
> > index 5c553c30d0..edfbfdca9e 100644
> > --- a/hw/char/serial-pci-multi.c
> > +++ b/hw/char/serial-pci-multi.c
> > @@ -181,10 +181,24 @@ static void multi_4x_serial_pci_class_initfn(ObjectClass *klass, void *data)
> >       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
> >   }
> >
> > +static void multi_serial_init(Object *o)
> > +{
> > +    PCIDevice *dev = PCI_DEVICE(o);
> > +    PCIMultiSerialState *pms = DO_UPCAST(PCIMultiSerialState, dev, dev);
> > +    size_t i, nports = multi_serial_get_port_count(PCI_DEVICE_GET_CLASS(dev));
> > +
> > +    for (i = 0; i < nports; i++) {
> > +        object_initialize_child(o, "serial[*]", &pms->state[i],
> > +                                sizeof(pms->state[i]),
> > +                                TYPE_SERIAL, &error_abort, NULL);
> > +    }
> > +}
> > +
> >   static const TypeInfo multi_2x_serial_pci_info = {
> >       .name          = "pci-serial-2x",
> >       .parent        = TYPE_PCI_DEVICE,
> >       .instance_size = sizeof(PCIMultiSerialState),
> > +    .instance_init = multi_serial_init,
> >       .class_init    = multi_2x_serial_pci_class_initfn,
> >       .interfaces = (InterfaceInfo[]) {
> >           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> > @@ -196,6 +210,7 @@ static const TypeInfo multi_4x_serial_pci_info = {
> >       .name          = "pci-serial-4x",
> >       .parent        = TYPE_PCI_DEVICE,
> >       .instance_size = sizeof(PCIMultiSerialState),
> > +    .instance_init = multi_serial_init,
> >       .class_init    = multi_4x_serial_pci_class_initfn,
> >       .interfaces = (InterfaceInfo[]) {
> >           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> > diff --git a/hw/char/serial-pci.c b/hw/char/serial-pci.c
> > index cb9b76e22b..f99b6c19e0 100644
> > --- a/hw/char/serial-pci.c
> > +++ b/hw/char/serial-pci.c
> > @@ -40,6 +40,8 @@ typedef struct PCISerialState {
> >       uint8_t prog_if;
> >   } PCISerialState;
> >
> > +#define TYPE_PCI_SERIAL "pci-serial"
> > +#define PCI_SERIAL(s) OBJECT_CHECK(PCISerialState, (s), TYPE_PCI_SERIAL)
> >
> >   static void serial_pci_realize(PCIDevice *dev, Error **errp)
> >   {
> > @@ -103,10 +105,19 @@ static void serial_pci_class_initfn(ObjectClass *klass, void *data)
> >       set_bit(DEVICE_CATEGORY_INPUT, dc->categories);
> >   }
> >
> > +static void serial_pci_init(Object *o)
> > +{
> > +    PCISerialState *ps = PCI_SERIAL(o);
> > +
> > +    object_initialize_child(o, "serial", &ps->state, sizeof(ps->state),
> > +                            TYPE_SERIAL, &error_abort, NULL);
> > +}
> > +
> >   static const TypeInfo serial_pci_info = {
> > -    .name          = "pci-serial",
> > +    .name          = TYPE_PCI_SERIAL,
> >       .parent        = TYPE_PCI_DEVICE,
> >       .instance_size = sizeof(PCISerialState),
> > +    .instance_init = serial_pci_init,
> >       .class_init    = serial_pci_class_initfn,
> >       .interfaces = (InterfaceInfo[]) {
> >           { INTERFACE_CONVENTIONAL_PCI_DEVICE },
> > diff --git a/hw/char/serial.c b/hw/char/serial.c
> > index b4aa250950..e0a5bec290 100644
> > --- a/hw/char/serial.c
> > +++ b/hw/char/serial.c
> > @@ -983,9 +983,8 @@ const MemoryRegionOps serial_io_ops = {
> >   SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> >                            Chardev *chr, MemoryRegion *system_io)
> >   {
> > -    SerialState *s;
> > -
> > -    s = g_malloc0(sizeof(SerialState));
> > +    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
> > +    SerialState *s = SERIAL(dev);
> >
> >       s->irq = irq;
> >       s->baudbase = baudbase;
> > @@ -993,6 +992,7 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> >       serial_realize_core(s, &error_fatal);
> >
> >       vmstate_register(NULL, base, &vmstate_serial, s);
> > +    qdev_init_nofail(dev);
> >
> >       memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
> >       memory_region_add_subregion(system_io, base, &s->io);
> > @@ -1000,6 +1000,20 @@ SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> >       return s;
> >   }
> >
> > +static void serial_class_init(ObjectClass *klass, void *data)
> > +{
> > +    DeviceClass *dc = DEVICE_CLASS(klass);
> > +
> > +    dc->user_creatable = false;
>
> Why? Due to the "chardev" property?

No, because it's not user usable as such (it's an internal device for
serialio/serialmm and others). And making abstract wouldn't work
either.

>
> > +}
> > +
> > +static const TypeInfo serial_info = {
> > +    .name = TYPE_SERIAL,
> > +    .parent = TYPE_DEVICE,
> > +    .instance_size = sizeof(SerialState),
> > +    .class_init = serial_class_init,
> > +};
> > +
> >   /* Memory mapped interface */
> >   static uint64_t serial_mm_read(void *opaque, hwaddr addr,
> >                                  unsigned size)
> > @@ -1045,9 +1059,8 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
> >                               qemu_irq irq, int baudbase,
> >                               Chardev *chr, enum device_endian end)
> >   {
> > -    SerialState *s;
> > -
> > -    s = g_malloc0(sizeof(SerialState));
> > +    DeviceState *dev = DEVICE(object_new(TYPE_SERIAL));
> > +    SerialState *s = SERIAL(dev);
> >
> >       s->it_shift = it_shift;
> >       s->irq = irq;
> > @@ -1056,9 +1069,17 @@ SerialState *serial_mm_init(MemoryRegion *address_space,
> >
> >       serial_realize_core(s, &error_fatal);
> >       vmstate_register(NULL, base, &vmstate_serial, s);
> > +    qdev_init_nofail(dev);
> >
> >       memory_region_init_io(&s->io, NULL, &serial_mm_ops[end], s,
> >                             "serial", 8 << it_shift);
> >       memory_region_add_subregion(address_space, base, &s->io);
> >       return s;
> >   }
> > +
> > +static void serial_register_types(void)
> > +{
> > +    type_register_static(&serial_info);
> > +}
> > +
> > +type_init(serial_register_types)
> > diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> > index 8be3d8a4f9..180cc7c24e 100644
> > --- a/include/hw/char/serial.h
> > +++ b/include/hw/char/serial.h
> > @@ -30,10 +30,13 @@
> >   #include "exec/memory.h"
> >   #include "qemu/fifo8.h"
> >   #include "chardev/char.h"
> > +#include "hw/sysbus.h"
> >
> >   #define UART_FIFO_LENGTH    16      /* 16550A Fifo Length */
> >
> >   typedef struct SerialState {
> > +    DeviceState parent;
> > +
> >       uint16_t divider;
> >       uint8_t rbr; /* receive register */
> >       uint8_t thr; /* transmit holding register */
> > @@ -84,7 +87,9 @@ void serial_realize_core(SerialState *s, Error **errp);
> >   void serial_exit_core(SerialState *s);
> >   void serial_set_frequency(SerialState *s, uint32_t frequency);
> >
> > -/* legacy pre qom */
> > +#define TYPE_SERIAL "serial"
> > +#define SERIAL(s) OBJECT_CHECK(SerialState, (s), TYPE_SERIAL)
> > +
> >   SerialState *serial_init(int base, qemu_irq irq, int baudbase,
> >                            Chardev *chr, MemoryRegion *system_io);
> >   SerialState *serial_mm_init(MemoryRegion *address_space,
> >
>
>


-- 
Marc-André Lureau

Re: [PATCH v4 28/37] qdev: use g_strcmp0() instead of open-coding it

[Eduardo Habkost]

On Wed, Nov 20, 2019 at 05:25:47PM +0100, Philippe Mathieu-Daudé wrote:
> On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> > Minor code simplification.
> > 
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >   hw/core/qdev.c | 7 ++-----
> >   1 file changed, 2 insertions(+), 5 deletions(-)
> > 
> > diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> > index cf1ba28fe3..c79befc865 100644
> > --- a/hw/core/qdev.c
> > +++ b/hw/core/qdev.c
> > @@ -394,11 +394,8 @@ static NamedGPIOList *qdev_get_named_gpio_list(DeviceState *dev,
> >       NamedGPIOList *ngl;
> >       QLIST_FOREACH(ngl, &dev->gpios, node) {
> > -        /* NULL is a valid and matchable name, otherwise do a normal
> > -         * strcmp match.
> > -         */
> > -        if ((!ngl->name && !name) ||
> > -                (name && ngl->name && strcmp(name, ngl->name) == 0)) {
> > +        /* NULL is a valid and matchable name. */
> > +        if (g_strcmp0(name, ngl->name) == 0) {
> 
> Cocci pattern?

This seems tricky because there are many different ways of
writing an equivalent expression.  I tried to catch a few
variations, but the script below didn't find any other
occurrences except for the one being touched by this patch.

@@
expression A, B;
@@
(
-(!A && !B) || (A && B && !strcmp(A, B))
+!g_strcmp0(A, B)
|
-(!B && !A) || (A && B && !strcmp(A, B))
+!g_strcmp0(A, B)
|
-(!A && !B) || (B && A && !strcmp(A, B))
+!g_strcmp0(A, B)
|
-(!B && !A) || (B && A && !strcmp(A, B))
+!g_strcmp0(A, B)
|
-(!A && !B) || (B && A && strcmp(A, B) == 0)
+!g_strcmp0(A, B)
|
-(!B && !A) || (B && A && strcmp(A, B) == 0)
+!g_strcmp0(A, B)
|
-(!A && !B) || (A && B && strcmp(A, B) == 0)
+!g_strcmp0(A, B)
|
-(!B && !A) || (A && B && strcmp(A, B) == 0)
+!g_strcmp0(A, B)
)

-- 
Eduardo

Re: [PATCH v4 25/37] dp8393x: replace PROP_PTR with PROP_LINK

[Laurent Vivier]

Le 20/11/2019 à 16:24, Marc-André Lureau a écrit :
> Link property is the correct way to pass a MemoryRegion to a device
> for DMA purposes.
> 
> Sidenote: as a sysbus device, this remains non-usercreatable
> even though we can drop the specific flag here.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  hw/m68k/q800.c      | 3 ++-
>  hw/mips/mips_jazz.c | 3 ++-
>  hw/net/dp8393x.c    | 7 +++----
>  3 files changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/hw/m68k/q800.c b/hw/m68k/q800.c
> index 4ca8678007..8f3eb6bfe7 100644
> --- a/hw/m68k/q800.c
> +++ b/hw/m68k/q800.c
> @@ -239,7 +239,8 @@ static void q800_init(MachineState *machine)
>      qdev_set_nic_properties(dev, &nd_table[0]);
>      qdev_prop_set_uint8(dev, "it_shift", 2);
>      qdev_prop_set_bit(dev, "big_endian", true);
> -    qdev_prop_set_ptr(dev, "dma_mr", get_system_memory());
> +    object_property_set_link(OBJECT(dev), OBJECT(get_system_memory()),
> +                             "dma_mr", &error_abort);
>      qdev_init_nofail(dev);
>      sysbus = SYS_BUS_DEVICE(dev);
>      sysbus_mmio_map(sysbus, 0, SONIC_BASE);
> diff --git a/hw/mips/mips_jazz.c b/hw/mips/mips_jazz.c
> index d978bb64a0..1518eb5e55 100644
> --- a/hw/mips/mips_jazz.c
> +++ b/hw/mips/mips_jazz.c
> @@ -284,7 +284,8 @@ static void mips_jazz_init(MachineState *machine,
>              dev = qdev_create(NULL, "dp8393x");
>              qdev_set_nic_properties(dev, nd);
>              qdev_prop_set_uint8(dev, "it_shift", 2);
> -            qdev_prop_set_ptr(dev, "dma_mr", rc4030_dma_mr);
> +            object_property_set_link(OBJECT(dev), OBJECT(rc4030_dma_mr),
> +                                     "dma_mr", &error_abort);
>              qdev_init_nofail(dev);
>              sysbus = SYS_BUS_DEVICE(dev);
>              sysbus_mmio_map(sysbus, 0, 0x80001000);
> diff --git a/hw/net/dp8393x.c b/hw/net/dp8393x.c
> index 3d991af163..cdc2631c0c 100644
> --- a/hw/net/dp8393x.c
> +++ b/hw/net/dp8393x.c
> @@ -175,7 +175,7 @@ typedef struct dp8393xState {
>      int loopback_packet;
>  
>      /* Memory access */
> -    void *dma_mr;
> +    MemoryRegion *dma_mr;
>      AddressSpace as;
>  } dp8393xState;
>  
> @@ -948,7 +948,8 @@ static const VMStateDescription vmstate_dp8393x = {
>  
>  static Property dp8393x_properties[] = {
>      DEFINE_NIC_PROPERTIES(dp8393xState, conf),
> -    DEFINE_PROP_PTR("dma_mr", dp8393xState, dma_mr),
> +    DEFINE_PROP_LINK("dma_mr", dp8393xState, dma_mr,
> +                     TYPE_MEMORY_REGION, MemoryRegion *),
>      DEFINE_PROP_UINT8("it_shift", dp8393xState, it_shift, 0),
>      DEFINE_PROP_BOOL("big_endian", dp8393xState, big_endian, false),
>      DEFINE_PROP_END_OF_LIST(),
> @@ -963,8 +964,6 @@ static void dp8393x_class_init(ObjectClass *klass, void *data)
>      dc->reset = dp8393x_reset;
>      dc->vmsd = &vmstate_dp8393x;
>      dc->props = dp8393x_properties;
> -    /* Reason: dma_mr property can't be set */
> -    dc->user_creatable = false;
>  }
>  
>  static const TypeInfo dp8393x_info = {
> 

Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Tested-by: Laurent Vivier <laurent@vivier.eu>

Re: [PATCH v4 06/37] serial: initial qom-ification

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:25, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Make SerialState a device (the following patches will introduce IO/MM
> sysbus serial devices)
>
> None of the serial_{,mm}_init() callers actually free the returned
> value (even if they did, it would be quite harmless), so we can change
> the object allocation at will.
>
> However, the devices that embed SerialState must now have their field
> QOM-initialized manually (isa, pci, pci-multi).
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>

> +static void serial_class_init(ObjectClass *klass, void *data)
> +{
> +    DeviceClass *dc = DEVICE_CLASS(klass);
> +
> +    dc->user_creatable = false;

The comment describing user_creatable in qdev-core.h says:
     * It should never be cleared without a comment explaining why it
     * is cleared.

So we should have a comment here.

Otherwise
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:28, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/mips/mips_mipssim.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index bfafa4d7e9..3cd0e6eb33 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>      if (serial_hd(0)) {
>          DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>
> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>          qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>          qdev_init_nofail(dev);
> --
> 2.24.0

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH v4 19/37] mips: use sysbus_add_io()

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:28, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/mips/mips_mipssim.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 3cd0e6eb33..2c2c7f25b2 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -227,8 +227,7 @@ mips_mipssim_init(MachineState *machine)
>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>          qdev_init_nofail(dev);
>          sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> -        memory_region_add_subregion(get_system_io(), 0x3f8,
> -                                    &SERIAL_IO(dev)->serial.io);
> +        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->serial.io);
>      }
>
>      if (nd_table[0].used)
> --

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Memory mapped serial device is in fact a sysbus device. The following
> patches will make use of sysbus facilities for resource and
> registration. In particular, "serial-mm: use sysbus facilities" will
> move internal serial realization to serial_mm_realize callback to
> follow qdev best practices.

What goes wrong if you just put the realize of smm->serial in
the right place to start with ?

thanks
-- PMM

Re: [PATCH v4 20/37] mips: use sysbus_mmio_get_region() instead of internal fields

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:28, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Register the memory region with sysbus_init_mmio() and look it up with
> sysbus_mmio_get_region() to avoid accessing internal device fields.
>
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/char/serial.c       | 1 +
>  hw/mips/mips_mipssim.c | 3 ++-

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

thanks
-- PMM

Re: [PATCH v4 27/37] leon3: use qdev gpio facilities for the PIL

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:30, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/sparc/leon3.c   | 6 ++++--
>  target/sparc/cpu.h | 1 -
>  2 files changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/hw/sparc/leon3.c b/hw/sparc/leon3.c
> index cac987373e..1a89d44e57 100644
> --- a/hw/sparc/leon3.c
> +++ b/hw/sparc/leon3.c
> @@ -230,8 +230,10 @@ static void leon3_generic_hw_init(MachineState *machine)
>
>      /* Allocate IRQ manager */
>      dev = qdev_create(NULL, TYPE_GRLIB_IRQMP);
> -    env->pil_irq = qemu_allocate_irq(leon3_set_pil_in, env, 0);
> -    qdev_connect_gpio_out_named(dev, "grlib-irq", 0, env->pil_irq);
> +    qdev_init_gpio_in_named_with_opaque(DEVICE(env), leon3_set_pil_in,
> +                                        env, "pil", 1);
> +    qdev_connect_gpio_out_named(dev, "grlib-irq", 0,
> +                                qdev_get_gpio_in_named(DEVICE(env), "pil", 0));
>      qdev_init_nofail(dev);
>      sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, LEON3_IRQMP_OFFSET);
>      env->irq_manager = dev;

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

Creating a gpio pin on some object that isn't yourself
looks a bit odd, but all this leon3 code is modifying
the CPU object from the outside anyway. Someday we might
tidy it up, but not today.

thanks
-- PMM

Re: [PATCH v4 21/37] sm501: make SerialMM a child, export chardev property

[Peter Maydell]

On Wed, 20 Nov 2019 at 15:31, Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Embed the SerialMM sybus device, and re-export its "chardev" property.
> That way, we can get rid of PROP_PTR "chr-state" and better track
> devices relationship.
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/display/sm501.c | 33 ++++++++++++++++++++++++---------
>  hw/sh4/r2d.c       |  2 +-
>  2 files changed, 25 insertions(+), 10 deletions(-)
>
> diff --git a/hw/display/sm501.c b/hw/display/sm501.c
> index 1f33c87e65..c4445b28f9 100644
> --- a/hw/display/sm501.c
> +++ b/hw/display/sm501.c
> @@ -1930,13 +1930,14 @@ typedef struct {
>      SM501State state;
>      uint32_t vram_size;
>      uint32_t base;
> -    void *chr_state;
> +    SerialMM serial;
>  } SM501SysBusState;
>
>  static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
>  {
>      SM501SysBusState *s = SYSBUS_SM501(dev);
>      SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
> +    SerialState *ss = &s->serial.serial;
>      DeviceState *usb_dev;
>
>      sm501_init(&s->state, dev, s->vram_size);
> @@ -1958,17 +1959,19 @@ static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
>      sysbus_pass_irq(sbd, SYS_BUS_DEVICE(usb_dev));
>
>      /* bridge to serial emulation module */
> -    if (s->chr_state) {
> -        serial_mm_init(&s->state.mmio_region, SM501_UART0, 2,
> -                       NULL, /* TODO : chain irq to IRL */
> -                       115200, s->chr_state, DEVICE_LITTLE_ENDIAN);
> +    /* FIXME: SM501_UART0 is always mapped, no need to check for the backend */
> +    if (qemu_chr_fe_backend_connected(&ss->chr)) {
> +        MemoryRegion *mr;
> +        qdev_init_nofail(DEVICE(&s->serial));
> +        mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->serial), 0);
> +        memory_region_add_subregion(&s->state.mmio_region, SM501_UART0, mr);
> +        /* TODO : chain irq to IRL */
>      }

I don't really understand what the FIXME is trying to
tell me here. If we don't need to check for the backend,
why is the code checking for it ? It means we have to fish
around inside the SerialMM's implementation, which seems odd.
Only mapping the UART registers if there happens to be a backend
connected also doesn't conceptually seem like the right behaviour,
because the registers should always exist. Since commit
12051d82f004024d5d the chardev mid-layer has correctly handled
the backend not being connected (ie having a NULL chardev),
so there's no longer any need for board/device code to special
case the lack of a chardev.

>  }
>
>  static Property sm501_sysbus_properties[] = {
>      DEFINE_PROP_UINT32("vram-size", SM501SysBusState, vram_size, 0),
>      DEFINE_PROP_UINT32("base", SM501SysBusState, base, 0),
> -    DEFINE_PROP_PTR("chr-state", SM501SysBusState, chr_state),
>      DEFINE_PROP_END_OF_LIST(),
>  };
>
> @@ -1999,9 +2002,20 @@ static void sm501_sysbus_class_init(ObjectClass *klass, void *data)
>      dc->props = sm501_sysbus_properties;
>      dc->reset = sm501_reset_sysbus;
>      dc->vmsd = &vmstate_sm501_sysbus;
> -    /* Note: pointer property "chr-state" may remain null, thus
> -     * no need for dc->user_creatable = false;
> -     */
> +}
> +
> +static void sm501_sysbus_init(Object *o)
> +{
> +    SM501SysBusState *sm501 = SYSBUS_SM501(o);
> +    SerialMM *smm = &sm501->serial;
> +
> +    sysbus_init_child_obj(o, "serial", smm, sizeof(SerialMM), TYPE_SERIAL_MM);
> +    qdev_set_legacy_instance_id(DEVICE(smm), SM501_UART0, 2);

The only board we use the sm501 sysbus device is the sh4 r2d
board, and we don't care about migration compatibility there
(indeed I would be unsurprised to find that it doesn't even work ;-))
So I think we can reasonably not set the legacy-instance-ID
and just declare in the commit message that this is a migration
compat break for that board.

> +    qdev_prop_set_uint8(DEVICE(smm), "regshift", 2);
> +    qdev_prop_set_uint8(DEVICE(smm), "endianness", DEVICE_LITTLE_ENDIAN);
> +
> +    object_property_add_alias(o, "chardev",
> +                              OBJECT(smm), "chardev", &error_abort);
>  }
>
>  static const TypeInfo sm501_sysbus_info = {
> @@ -2009,6 +2023,7 @@ static const TypeInfo sm501_sysbus_info = {
>      .parent        = TYPE_SYS_BUS_DEVICE,
>      .instance_size = sizeof(SM501SysBusState),
>      .class_init    = sm501_sysbus_class_init,
> +    .instance_init = sm501_sysbus_init,
>  };
>
>  #define TYPE_PCI_SM501 "sm501"
> diff --git a/hw/sh4/r2d.c b/hw/sh4/r2d.c
> index ee0840f380..72bb5285cc 100644
> --- a/hw/sh4/r2d.c
> +++ b/hw/sh4/r2d.c
> @@ -272,7 +272,7 @@ static void r2d_init(MachineState *machine)
>      busdev = SYS_BUS_DEVICE(dev);
>      qdev_prop_set_uint32(dev, "vram-size", SM501_VRAM_SIZE);
>      qdev_prop_set_uint32(dev, "base", 0x10000000);
> -    qdev_prop_set_ptr(dev, "chr-state", serial_hd(2));
> +    qdev_prop_set_chr(dev, "chardev", serial_hd(2));
>      qdev_init_nofail(dev);
>      sysbus_mmio_map(busdev, 0, 0x10000000);
>      sysbus_mmio_map(busdev, 1, 0x13e00000);
> --
> 2.24.0

thanks
-- PMM

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Marc-André Lureau]

On Thu, Nov 21, 2019 at 5:47 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
> <marcandre.lureau@redhat.com> wrote:
> >
> > Memory mapped serial device is in fact a sysbus device. The following
> > patches will make use of sysbus facilities for resource and
> > registration. In particular, "serial-mm: use sysbus facilities" will
> > move internal serial realization to serial_mm_realize callback to
> > follow qdev best practices.
>
> What goes wrong if you just put the realize of smm->serial in
> the right place to start with ?

You mean squash the following patches?
Sometime I go too fast, sometime it's too slow.

Decide what you prefer, this doesnt' matter much to me.


-- 
Marc-André Lureau

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Peter Maydell]

On Thu, 21 Nov 2019 at 18:15, Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
> On Thu, Nov 21, 2019 at 5:47 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> >
> > On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
> > <marcandre.lureau@redhat.com> wrote:
> > >
> > > Memory mapped serial device is in fact a sysbus device. The following
> > > patches will make use of sysbus facilities for resource and
> > > registration. In particular, "serial-mm: use sysbus facilities" will
> > > move internal serial realization to serial_mm_realize callback to
> > > follow qdev best practices.
> >
> > What goes wrong if you just put the realize of smm->serial in
> > the right place to start with ?
>
> You mean squash the following patches?

No, I meant just having this patch have

static void serial_mm_realize(DeviceState *dev, Error **errp)
{
    SerialMM *smm = SERIAL_MM(dev);
    SerialState *s = &smm->serial;

    object_property_set_bool(OBJECT(dev), true, "realized", &err);
    if (err) {
        error_propagate(errp, err);
        return;
    }
}

and

+ dc->realize = serial_mm_realize;

rather than manually doing the qdev_init_nofail()
in serial_mm_init(). This seems to me like an integral
part of the change to doing the init of the subdevice in the
init method, so it would be better unless there's a reason
why it breaks something. The rest of patch 15 (which is
what currently makes the equivalent change to realize) is all
about passing through the properties and exposing the
sysbus MMIO/irq regions and should stay a separate patch.

(setting the 'realized' property is better in a realize method
than using qdev_init_nofail() because it means we can propagate
any error outward rather than killing qemu.)

thanks
-- PMM

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Marc-André Lureau]

Hi

On Thu, Nov 21, 2019 at 10:24 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Thu, 21 Nov 2019 at 18:15, Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
> >
> > On Thu, Nov 21, 2019 at 5:47 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > >
> > > On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
> > > <marcandre.lureau@redhat.com> wrote:
> > > >
> > > > Memory mapped serial device is in fact a sysbus device. The following
> > > > patches will make use of sysbus facilities for resource and
> > > > registration. In particular, "serial-mm: use sysbus facilities" will
> > > > move internal serial realization to serial_mm_realize callback to
> > > > follow qdev best practices.
> > >
> > > What goes wrong if you just put the realize of smm->serial in
> > > the right place to start with ?
> >
> > You mean squash the following patches?
>
> No, I meant just having this patch have
>
> static void serial_mm_realize(DeviceState *dev, Error **errp)
> {
>     SerialMM *smm = SERIAL_MM(dev);
>     SerialState *s = &smm->serial;
>
>     object_property_set_bool(OBJECT(dev), true, "realized", &err);
>     if (err) {
>         error_propagate(errp, err);
>         return;
>     }
> }
>
> and
>
> + dc->realize = serial_mm_realize;
>
> rather than manually doing the qdev_init_nofail()
> in serial_mm_init(). This seems to me like an integral
> part of the change to doing the init of the subdevice in the
> init method, so it would be better unless there's a reason
> why it breaks something. The rest of patch 15 (which is
> what currently makes the equivalent change to realize) is all
> about passing through the properties and exposing the
> sysbus MMIO/irq regions and should stay a separate patch.
>
> (setting the 'realized' property is better in a realize method
> than using qdev_init_nofail() because it means we can propagate
> any error outward rather than killing qemu.)

Ok, but I implemented realize and moved serial init in "serial: make
SerialIO a sysbus device".

I propose to add another patch to follow your suggestion to use
set_boot("realize", true, errp) on top.

-- 
Marc-André Lureau

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Corey Minyard]

On Wed, Nov 20, 2019 at 07:24:36PM +0400, Marc-André Lureau wrote:
> Instead, set the initial data field directly.
> 
> (the initial data is an array of 256 bytes. As I don't know if it may
> change over time, I keep the pointer to original buffer as is, but it
> might be worth to consider to copy it instead)

This looks ok to me, I can add it to my tree.  I think we are in hard
freeze now, so this will have to wait until that's done.

-corey

> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/i2c/smbus_eeprom.c | 17 ++++++++---------
>  1 file changed, 8 insertions(+), 9 deletions(-)
> 
> diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
> index 54c86a0112..533c728b3b 100644
> --- a/hw/i2c/smbus_eeprom.c
> +++ b/hw/i2c/smbus_eeprom.c
> @@ -44,7 +44,7 @@
>  typedef struct SMBusEEPROMDevice {
>      SMBusDevice smbusdev;
>      uint8_t data[SMBUS_EEPROM_SIZE];
> -    void *init_data;
> +    uint8_t *init_data;
>      uint8_t offset;
>      bool accessed;
>  } SMBusEEPROMDevice;
> @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
>  
>  static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
>  {
> +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
> +
>      smbus_eeprom_reset(dev);
> +    if (eeprom->init_data == NULL) {
> +        error_setg(errp, "init_data cannot be NULL");
> +    }
>  }
>  
> -static Property smbus_eeprom_properties[] = {
> -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
> -    DEFINE_PROP_END_OF_LIST(),
> -};
> -
>  static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>  {
>      DeviceClass *dc = DEVICE_CLASS(klass);
> @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>      dc->reset = smbus_eeprom_reset;
>      sc->receive_byte = eeprom_receive_byte;
>      sc->write_data = eeprom_write_data;
> -    dc->props = smbus_eeprom_properties;
>      dc->vmsd = &vmstate_smbus_eeprom;
> -    /* Reason: pointer property "data" */
> +    /* Reason: init_data */
>      dc->user_creatable = false;
>  }
>  
> @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
>  
>      dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
>      qdev_prop_set_uint8(dev, "address", address);
> -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
> +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
>      qdev_init_nofail(dev);
>  }
>  
> -- 
> 2.24.0
> 

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Peter Maydell]

On Thu, 21 Nov 2019 at 18:51, Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
> Hi
>
> On Thu, Nov 21, 2019 at 10:24 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> >
> > On Thu, 21 Nov 2019 at 18:15, Marc-André Lureau
> > <marcandre.lureau@gmail.com> wrote:
> > >
> > > On Thu, Nov 21, 2019 at 5:47 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > > >
> > > > On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
> > > > <marcandre.lureau@redhat.com> wrote:
> > > > >
> > > > > Memory mapped serial device is in fact a sysbus device. The following
> > > > > patches will make use of sysbus facilities for resource and
> > > > > registration. In particular, "serial-mm: use sysbus facilities" will
> > > > > move internal serial realization to serial_mm_realize callback to
> > > > > follow qdev best practices.
> > > >
> > > > What goes wrong if you just put the realize of smm->serial in
> > > > the right place to start with ?
> > >
> > > You mean squash the following patches?
> >
> > No, I meant just having this patch have
> >
> > static void serial_mm_realize(DeviceState *dev, Error **errp)
> > {
> >     SerialMM *smm = SERIAL_MM(dev);
> >     SerialState *s = &smm->serial;
> >
> >     object_property_set_bool(OBJECT(dev), true, "realized", &err);
> >     if (err) {
> >         error_propagate(errp, err);
> >         return;
> >     }
> > }
> >
> > and
> >
> > + dc->realize = serial_mm_realize;
> >
> > rather than manually doing the qdev_init_nofail()
> > in serial_mm_init(). This seems to me like an integral
> > part of the change to doing the init of the subdevice in the
> > init method, so it would be better unless there's a reason
> > why it breaks something. The rest of patch 15 (which is
> > what currently makes the equivalent change to realize) is all
> > about passing through the properties and exposing the
> > sysbus MMIO/irq regions and should stay a separate patch.
> >
> > (setting the 'realized' property is better in a realize method
> > than using qdev_init_nofail() because it means we can propagate
> > any error outward rather than killing qemu.)
>
> Ok, but I implemented realize and moved serial init in "serial: make
> SerialIO a sysbus device".

That patch is for the TYPE_SERIAL_IO device, isn't it? It
changes both serial_io_instance_init and serial_io_realize
to do the instance init and realize of the TYPE_SERIAL device,
so it doesn't have the same "only doing half a change" issue
that this patch has for TYPE_SERIAL_MM.

thanks
-- PMM

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Instead, set the initial data field directly.
> 
> (the initial data is an array of 256 bytes. As I don't know if it may
> change over time, I keep the pointer to original buffer as is, but it
> might be worth to consider to copy it instead)
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/i2c/smbus_eeprom.c | 17 ++++++++---------
>   1 file changed, 8 insertions(+), 9 deletions(-)
> 
> diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
> index 54c86a0112..533c728b3b 100644
> --- a/hw/i2c/smbus_eeprom.c
> +++ b/hw/i2c/smbus_eeprom.c
> @@ -44,7 +44,7 @@
>   typedef struct SMBusEEPROMDevice {
>       SMBusDevice smbusdev;
>       uint8_t data[SMBUS_EEPROM_SIZE];
> -    void *init_data;
> +    uint8_t *init_data;
>       uint8_t offset;
>       bool accessed;
>   } SMBusEEPROMDevice;
> @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
>   
>   static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
>   {
> +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
> +
>       smbus_eeprom_reset(dev);
> +    if (eeprom->init_data == NULL) {
> +        error_setg(errp, "init_data cannot be NULL");
> +    }
>   }
>   
> -static Property smbus_eeprom_properties[] = {
> -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
> -    DEFINE_PROP_END_OF_LIST(),
> -};
> -
>   static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>   {
>       DeviceClass *dc = DEVICE_CLASS(klass);
> @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>       dc->reset = smbus_eeprom_reset;
>       sc->receive_byte = eeprom_receive_byte;
>       sc->write_data = eeprom_write_data;
> -    dc->props = smbus_eeprom_properties;
>       dc->vmsd = &vmstate_smbus_eeprom;
> -    /* Reason: pointer property "data" */
> +    /* Reason: init_data */
>       dc->user_creatable = false;
>   }
>   
> @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
>   
>       dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
>       qdev_prop_set_uint8(dev, "address", address);
> -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
> +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;

Isn't this a QOM anti-pattern?

>       qdev_init_nofail(dev);
>   }
>   
> 

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Marc-André Lureau]

On Fri, Nov 22, 2019 at 3:07 PM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> > Instead, set the initial data field directly.
> >
> > (the initial data is an array of 256 bytes. As I don't know if it may
> > change over time, I keep the pointer to original buffer as is, but it
> > might be worth to consider to copy it instead)
> >
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >   hw/i2c/smbus_eeprom.c | 17 ++++++++---------
> >   1 file changed, 8 insertions(+), 9 deletions(-)
> >
> > diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
> > index 54c86a0112..533c728b3b 100644
> > --- a/hw/i2c/smbus_eeprom.c
> > +++ b/hw/i2c/smbus_eeprom.c
> > @@ -44,7 +44,7 @@
> >   typedef struct SMBusEEPROMDevice {
> >       SMBusDevice smbusdev;
> >       uint8_t data[SMBUS_EEPROM_SIZE];
> > -    void *init_data;
> > +    uint8_t *init_data;
> >       uint8_t offset;
> >       bool accessed;
> >   } SMBusEEPROMDevice;
> > @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
> >
> >   static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
> >   {
> > +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
> > +
> >       smbus_eeprom_reset(dev);
> > +    if (eeprom->init_data == NULL) {
> > +        error_setg(errp, "init_data cannot be NULL");
> > +    }
> >   }
> >
> > -static Property smbus_eeprom_properties[] = {
> > -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
> > -    DEFINE_PROP_END_OF_LIST(),
> > -};
> > -
> >   static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> >   {
> >       DeviceClass *dc = DEVICE_CLASS(klass);
> > @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> >       dc->reset = smbus_eeprom_reset;
> >       sc->receive_byte = eeprom_receive_byte;
> >       sc->write_data = eeprom_write_data;
> > -    dc->props = smbus_eeprom_properties;
> >       dc->vmsd = &vmstate_smbus_eeprom;
> > -    /* Reason: pointer property "data" */
> > +    /* Reason: init_data */
> >       dc->user_creatable = false;
> >   }
> >
> > @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
> >
> >       dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
> >       qdev_prop_set_uint8(dev, "address", address);
> > -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
> > +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
>
> Isn't this a QOM anti-pattern?

Sort of, but using prop_ptr in the first place is an anti-pattern. So
I am not improving the situation much here, but just getting rid of
prop_ptr usage.

Using an array of bytes instead could be done in a follow-up patch. I
can leave a fixme if that helps.

>
> >       qdev_init_nofail(dev);
> >   }
> >
> >
>

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Philippe Mathieu-Daudé]

On 11/22/19 12:12 PM, Marc-André Lureau wrote:
> On Fri, Nov 22, 2019 at 3:07 PM Philippe Mathieu-Daudé
> <philmd@redhat.com> wrote:
>>
>> On 11/20/19 4:24 PM, Marc-André Lureau wrote:
>>> Instead, set the initial data field directly.
>>>
>>> (the initial data is an array of 256 bytes. As I don't know if it may
>>> change over time, I keep the pointer to original buffer as is, but it
>>> might be worth to consider to copy it instead)
>>>
>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>> ---
>>>    hw/i2c/smbus_eeprom.c | 17 ++++++++---------
>>>    1 file changed, 8 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
>>> index 54c86a0112..533c728b3b 100644
>>> --- a/hw/i2c/smbus_eeprom.c
>>> +++ b/hw/i2c/smbus_eeprom.c
>>> @@ -44,7 +44,7 @@
>>>    typedef struct SMBusEEPROMDevice {
>>>        SMBusDevice smbusdev;
>>>        uint8_t data[SMBUS_EEPROM_SIZE];
>>> -    void *init_data;
>>> +    uint8_t *init_data;
>>>        uint8_t offset;
>>>        bool accessed;
>>>    } SMBusEEPROMDevice;
>>> @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
>>>
>>>    static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
>>>    {
>>> +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
>>> +
>>>        smbus_eeprom_reset(dev);
>>> +    if (eeprom->init_data == NULL) {
>>> +        error_setg(errp, "init_data cannot be NULL");
>>> +    }
>>>    }
>>>
>>> -static Property smbus_eeprom_properties[] = {
>>> -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
>>> -    DEFINE_PROP_END_OF_LIST(),
>>> -};
>>> -
>>>    static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>>>    {
>>>        DeviceClass *dc = DEVICE_CLASS(klass);
>>> @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
>>>        dc->reset = smbus_eeprom_reset;
>>>        sc->receive_byte = eeprom_receive_byte;
>>>        sc->write_data = eeprom_write_data;
>>> -    dc->props = smbus_eeprom_properties;
>>>        dc->vmsd = &vmstate_smbus_eeprom;
>>> -    /* Reason: pointer property "data" */
>>> +    /* Reason: init_data */
>>>        dc->user_creatable = false;
>>>    }
>>>
>>> @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
>>>
>>>        dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
>>>        qdev_prop_set_uint8(dev, "address", address);
>>> -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
>>> +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
>>
>> Isn't this a QOM anti-pattern?
> 
> Sort of, but using prop_ptr in the first place is an anti-pattern. So
> I am not improving the situation much here, but just getting rid of
> prop_ptr usage.

Alright.

> Using an array of bytes instead could be done in a follow-up patch. I
> can leave a fixme if that helps.

IIRC there was a discussion earlier this year about using block backend 
for this device.

>>
>>>        qdev_init_nofail(dev);
>>>    }
>>>
>>>
>>
> 

Re: [PATCH v4 12/37] serial: start making SerialMM a sysbus device

[Marc-André Lureau]

On Fri, Nov 22, 2019 at 2:11 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Thu, 21 Nov 2019 at 18:51, Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
> >
> > Hi
> >
> > On Thu, Nov 21, 2019 at 10:24 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > >
> > > On Thu, 21 Nov 2019 at 18:15, Marc-André Lureau
> > > <marcandre.lureau@gmail.com> wrote:
> > > >
> > > > On Thu, Nov 21, 2019 at 5:47 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > > > >
> > > > > On Wed, 20 Nov 2019 at 15:27, Marc-André Lureau
> > > > > <marcandre.lureau@redhat.com> wrote:
> > > > > >
> > > > > > Memory mapped serial device is in fact a sysbus device. The following
> > > > > > patches will make use of sysbus facilities for resource and
> > > > > > registration. In particular, "serial-mm: use sysbus facilities" will
> > > > > > move internal serial realization to serial_mm_realize callback to
> > > > > > follow qdev best practices.
> > > > >
> > > > > What goes wrong if you just put the realize of smm->serial in
> > > > > the right place to start with ?
> > > >
> > > > You mean squash the following patches?
> > >
> > > No, I meant just having this patch have
> > >
> > > static void serial_mm_realize(DeviceState *dev, Error **errp)
> > > {
> > >     SerialMM *smm = SERIAL_MM(dev);
> > >     SerialState *s = &smm->serial;
> > >
> > >     object_property_set_bool(OBJECT(dev), true, "realized", &err);
> > >     if (err) {
> > >         error_propagate(errp, err);
> > >         return;
> > >     }
> > > }
> > >
> > > and
> > >
> > > + dc->realize = serial_mm_realize;
> > >
> > > rather than manually doing the qdev_init_nofail()
> > > in serial_mm_init(). This seems to me like an integral
> > > part of the change to doing the init of the subdevice in the
> > > init method, so it would be better unless there's a reason
> > > why it breaks something. The rest of patch 15 (which is
> > > what currently makes the equivalent change to realize) is all
> > > about passing through the properties and exposing the
> > > sysbus MMIO/irq regions and should stay a separate patch.
> > >
> > > (setting the 'realized' property is better in a realize method
> > > than using qdev_init_nofail() because it means we can propagate
> > > any error outward rather than killing qemu.)
> >
> > Ok, but I implemented realize and moved serial init in "serial: make
> > SerialIO a sysbus device".
>
> That patch is for the TYPE_SERIAL_IO device, isn't it? It
> changes both serial_io_instance_init and serial_io_realize
> to do the instance init and realize of the TYPE_SERIAL device,
> so it doesn't have the same "only doing half a change" issue
> that this patch has for TYPE_SERIAL_MM.
>

Ok, I got it, thanks.




-- 
Marc-André Lureau

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 1001 bytes --]

On Wednesday, November 20, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/mips/mips_mipssim.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index bfafa4d7e9..3cd0e6eb33 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>      if (serial_hd(0)) {
>          DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>
> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>          qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>          qdev_init_nofail(dev);
> --


Please mention in your commit message where the default baudbase is set.

Also, is there a guarantie that default value 115200 will never change in
future?

Yours, Aleksandar



> 2.24.0
>
>
>

[-- Attachment #2: Type: text/html, Size: 1534 bytes --]

Re: [PATCH v4 17/37] mips: inline serial_init()

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 3674 bytes --]

On Wednesday, November 20, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> The function is specific to mipssim, let's inline it.
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/char/serial.c         | 16 ----------------
>  hw/mips/mips_mipssim.c   | 15 ++++++++++++---
>  include/hw/char/serial.h |  2 --
>  3 files changed, 12 insertions(+), 21 deletions(-)
>
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 164146ede8..23f0b02516 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
>      .class_init = serial_io_class_init,
>  };
>
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                         Chardev *chr, MemoryRegion *system_io)
> -{
> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> -
> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> -    qdev_init_nofail(DEVICE(sio));
> -
> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
> -
> -    return sio;
> -}
> -
>  static Property serial_properties[] = {
>      DEFINE_PROP_CHR("chardev", SerialState, chr),
>      DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 282bbecb24..bfafa4d7e9 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -40,6 +40,7 @@
>  #include "hw/loader.h"
>  #include "elf.h"
>  #include "hw/sysbus.h"
> +#include "hw/qdev-properties.h"
>  #include "exec/address-spaces.h"
>  #include "qemu/error-report.h"
>  #include "sysemu/qtest.h"
> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
>       * A single 16450 sits at offset 0x3f8. It is attached to
>       * MIPS CPU INT2, which is interrupt 4.
>       */
> -    if (serial_hd(0))
> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
> -                    get_system_io());
> +    if (serial_hd(0)) {
> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> +
> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> +        qdev_init_nofail(dev);
> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> +        memory_region_add_subregion(get_system_io(), 0x3f8,
> +                                    &SERIAL_IO(dev)->serial.io);
> +    }



Please explain why the code in the deleted function and the new function
are not identical. Why is the new code better?

Thanks, Aleksandar


>
>      if (nd_table[0].used)
>          /* MIPSnet uses the MIPS CPU INT0, which is interrupt 2. */
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index d356ba838c..535fa23a2b 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -108,8 +108,6 @@ void serial_set_frequency(SerialState *s, uint32_t
> frequency);
>  #define TYPE_SERIAL_IO "serial-io"
>  #define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
>
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                      Chardev *chr, MemoryRegion *system_io);
>  SerialMM *serial_mm_init(MemoryRegion *address_space,
>                           hwaddr base, int regshift,
>                           qemu_irq irq, int baudbase,
> --
> 2.24.0
>
>
>

[-- Attachment #2: Type: text/html, Size: 4739 bytes --]

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Marc-André Lureau]

Hi

On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
<aleksandar.m.mail@gmail.com> wrote:
>
>
>
> On Wednesday, November 20, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
>>
>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>> ---
>>  hw/mips/mips_mipssim.c | 1 -
>>  1 file changed, 1 deletion(-)
>>
>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>> index bfafa4d7e9..3cd0e6eb33 100644
>> --- a/hw/mips/mips_mipssim.c
>> +++ b/hw/mips/mips_mipssim.c
>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>>      if (serial_hd(0)) {
>>          DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>>
>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>>          qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>>          qdev_init_nofail(dev);
>> --
>
>
> Please mention in your commit message where the default baudbase is set.

ok

> Also, is there a guarantie that default value 115200 will never change in future?

The level of stability on properties in general is unclear to me.

Given that 115200 is standard for serial, it is unlikely to change
though.. We can have an assert there instead?

Peter, what do you think? thanks

Re: [PATCH v4 17/37] mips: inline serial_init()

[Marc-André Lureau]

Hi

On Mon, Nov 25, 2019 at 2:12 PM Aleksandar Markovic
<aleksandar.m.mail@gmail.com> wrote:
>
>
>
> On Wednesday, November 20, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
>>
>> The function is specific to mipssim, let's inline it.
>>
>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>> ---
>>  hw/char/serial.c         | 16 ----------------
>>  hw/mips/mips_mipssim.c   | 15 ++++++++++++---
>>  include/hw/char/serial.h |  2 --
>>  3 files changed, 12 insertions(+), 21 deletions(-)
>>
>> diff --git a/hw/char/serial.c b/hw/char/serial.c
>> index 164146ede8..23f0b02516 100644
>> --- a/hw/char/serial.c
>> +++ b/hw/char/serial.c
>> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
>>      .class_init = serial_io_class_init,
>>  };
>>
>> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
>> -                         Chardev *chr, MemoryRegion *system_io)
>> -{
>> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
>> -
>> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
>> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
>> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
>> -    qdev_init_nofail(DEVICE(sio));
>> -
>> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
>> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
>> -
>> -    return sio;
>> -}
>> -
>>  static Property serial_properties[] = {
>>      DEFINE_PROP_CHR("chardev", SerialState, chr),
>>      DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>> index 282bbecb24..bfafa4d7e9 100644
>> --- a/hw/mips/mips_mipssim.c
>> +++ b/hw/mips/mips_mipssim.c
>> @@ -40,6 +40,7 @@
>>  #include "hw/loader.h"
>>  #include "elf.h"
>>  #include "hw/sysbus.h"
>> +#include "hw/qdev-properties.h"
>>  #include "exec/address-spaces.h"
>>  #include "qemu/error-report.h"
>>  #include "sysemu/qtest.h"
>> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
>>       * A single 16450 sits at offset 0x3f8. It is attached to
>>       * MIPS CPU INT2, which is interrupt 4.
>>       */
>> -    if (serial_hd(0)) a
>> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
>> -                    get_system_io());
>> +    if (serial_hd(0)) {
>> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>> +
>> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>> +        qdev_init_nofail(dev);
>> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
>> +        memory_region_add_subregion(get_system_io(), 0x3f8,
>> +                                    &SERIAL_IO(dev)->serial.io);
>> +    }
>
>
>
> Please explain why the code in the deleted function and the new function are not identical. Why is the new code better?

What differences do you spot?

serial_init() looks very generic, but it is not, and this is
confusing. QDev users should call qdev_create().

If there is a single user, it is better to inline the code where it is
being used, or move the function there.

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Philippe Mathieu-Daudé]

On 11/25/19 11:12 AM, Marc-André Lureau wrote:
> Hi
> 
> On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
> <aleksandar.m.mail@gmail.com> wrote:
>>
>>
>>
>> On Wednesday, November 20, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
>>>
>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>> ---
>>>   hw/mips/mips_mipssim.c | 1 -
>>>   1 file changed, 1 deletion(-)
>>>
>>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>>> index bfafa4d7e9..3cd0e6eb33 100644
>>> --- a/hw/mips/mips_mipssim.c
>>> +++ b/hw/mips/mips_mipssim.c
>>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>>>       if (serial_hd(0)) {
>>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>>>
>>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>>>           qdev_init_nofail(dev);
>>> --
>>
>>
>> Please mention in your commit message where the default baudbase is set.
> 
> ok
> 
>> Also, is there a guarantie that default value 115200 will never change in future?
> 
> The level of stability on properties in general is unclear to me.
> 
> Given that 115200 is standard for serial, it is unlikely to change
> though.. We can have an assert there instead?
> 
> Peter, what do you think? thanks

This property confused me by the past. It is _not_ the baudrate.
It is the input frequency clocking the UART ('XIN' pin, Xtal INput).

Each board has its own frequency, and it can even be variable (the clock 
domain tree can reconfigure it at a different rate).

I'm not sure it makes sense to have a default, and I don't know what is 
the frequency modeled by the SPIM simulator.

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Aleksandar Markovic]

On Mon, Nov 25, 2019 at 12:26 PM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/25/19 11:12 AM, Marc-André Lureau wrote:
> > Hi
> >
> > On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
> > <aleksandar.m.mail@gmail.com> wrote:
> >>
> >>
> >>
> >> On Wednesday, November 20, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
> >>>
> >>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> >>> ---
> >>>   hw/mips/mips_mipssim.c | 1 -
> >>>   1 file changed, 1 deletion(-)
> >>>
> >>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> >>> index bfafa4d7e9..3cd0e6eb33 100644
> >>> --- a/hw/mips/mips_mipssim.c
> >>> +++ b/hw/mips/mips_mipssim.c
> >>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
> >>>       if (serial_hd(0)) {
> >>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> >>>
> >>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> >>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> >>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> >>>           qdev_init_nofail(dev);
> >>> --
> >>
> >>
> >> Please mention in your commit message where the default baudbase is set.
> >
> > ok
> >
> >> Also, is there a guarantie that default value 115200 will never change in future?
> >
> > The level of stability on properties in general is unclear to me.
> >
> > Given that 115200 is standard for serial, it is unlikely to change
> > though.. We can have an assert there instead?
> >
> > Peter, what do you think? thanks
>
> This property confused me by the past. It is _not_ the baudrate.

The name is "boudbase" (whatever that means), so not "boudrate".

Can we perhaps track the "inventor" of the property?

Google search for the word "boudbase" gives me an address of a person
in Maryland, whose last name is Boudbase, and also another address of
apparently the same person in Fresno, CA. No serial-device-related
results occurred (within the first page of Google results, at least).

Sincerely,
Aleksandar

> It is the input frequency clocking the UART ('XIN' pin, Xtal INput).
>
> Each board has its own frequency, and it can even be variable (the clock
> domain tree can reconfigure it at a different rate).
>
> I'm not sure it makes sense to have a default, and I don't know what is
> the frequency modeled by the SPIM simulator.
>

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Philippe Mathieu-Daudé]

On 11/25/19 12:26 PM, Philippe Mathieu-Daudé wrote:
> On 11/25/19 11:12 AM, Marc-André Lureau wrote:
>> Hi
>>
>> On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
>> <aleksandar.m.mail@gmail.com> wrote:
>>>
>>>
>>>
>>> On Wednesday, November 20, 2019, Marc-André Lureau 
>>> <marcandre.lureau@redhat.com> wrote:
>>>>
>>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>>> ---
>>>>   hw/mips/mips_mipssim.c | 1 -
>>>>   1 file changed, 1 deletion(-)
>>>>
>>>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>>>> index bfafa4d7e9..3cd0e6eb33 100644
>>>> --- a/hw/mips/mips_mipssim.c
>>>> +++ b/hw/mips/mips_mipssim.c
>>>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>>>>       if (serial_hd(0)) {
>>>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>>>>
>>>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>>>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>>>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>>>>           qdev_init_nofail(dev);
>>>> -- 
>>>
>>>
>>> Please mention in your commit message where the default baudbase is set.
>>
>> ok
>>
>>> Also, is there a guarantie that default value 115200 will never 
>>> change in future?
>>
>> The level of stability on properties in general is unclear to me.
>>
>> Given that 115200 is standard for serial, it is unlikely to change
>> though.. We can have an assert there instead?
>>
>> Peter, what do you think? thanks
> 
> This property confused me by the past. It is _not_ the baudrate.
> It is the input frequency clocking the UART ('XIN' pin, Xtal INput).
> 
> Each board has its own frequency, and it can even be variable (the clock 
> domain tree can reconfigure it at a different rate).

Laurent pointed me to the following commit which confirms my interpretation:

$ git show 038eaf82c853
commit 038eaf82c853f3bf8d4c106c0677bbf4adada7de
Author: Stefan Weil <weil@mail.berlios.de>
Date:   Sat Oct 31 11:28:11 2009 +0100

     serial: Add interface to set reference oscillator frequency

     Many (most?) serial interfaces have a programmable
     clock which provides the reference frequency ("baudbase").
     So a fixed baudbase which is only set once can be wrong.

     omap1.c is an example which could use the new interface
     to change baudbase when the programmable clock changes.
     ar7 system emulation (still not part of standard QEMU)
     is similar to omap and already uses serial_set_frequency.

     Signed-off-by: Stefan Weil <weil@mail.berlios.de>
     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

diff --git a/hw/pc.h b/hw/pc.h
index 15fff8d103..03ffc91536 100644
--- a/hw/pc.h
+++ b/hw/pc.h
@@ -13,6 +13,7 @@ SerialState *serial_mm_init (target_phys_addr_t base, 
int it_shift,
                               qemu_irq irq, int baudbase,
                               CharDriverState *chr, int ioregister);
  SerialState *serial_isa_init(int index, CharDriverState *chr);
+void serial_set_frequency(SerialState *s, uint32_t frequency);

  /* parallel.c */

diff --git a/hw/serial.c b/hw/serial.c
index fa12dcc075..0063260569 100644
--- a/hw/serial.c
+++ b/hw/serial.c
@@ -730,6 +730,13 @@ static void serial_init_core(SerialState *s)
                            serial_event, s);
  }

+/* Change the main reference oscillator frequency. */
+void serial_set_frequency(SerialState *s, uint32_t frequency)
+{
+    s->baudbase = frequency;
+    serial_update_parameters(s);
+}
+

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Philippe Mathieu-Daudé]

On 11/25/19 1:54 PM, Philippe Mathieu-Daudé wrote:
> On 11/25/19 12:26 PM, Philippe Mathieu-Daudé wrote:
>> On 11/25/19 11:12 AM, Marc-André Lureau wrote:
>>> Hi
>>>
>>> On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
>>> <aleksandar.m.mail@gmail.com> wrote:
>>>>
>>>>
>>>>
>>>> On Wednesday, November 20, 2019, Marc-André Lureau 
>>>> <marcandre.lureau@redhat.com> wrote:
>>>>>
>>>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>>>> ---
>>>>>   hw/mips/mips_mipssim.c | 1 -
>>>>>   1 file changed, 1 deletion(-)
>>>>>
>>>>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>>>>> index bfafa4d7e9..3cd0e6eb33 100644
>>>>> --- a/hw/mips/mips_mipssim.c
>>>>> +++ b/hw/mips/mips_mipssim.c
>>>>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
>>>>>       if (serial_hd(0)) {
>>>>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>>>>>
>>>>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>>>>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>>>>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>>>>>           qdev_init_nofail(dev);
>>>>> -- 
>>>>
>>>>
>>>> Please mention in your commit message where the default baudbase is 
>>>> set.
>>>
>>> ok
>>>
>>>> Also, is there a guarantie that default value 115200 will never 
>>>> change in future?
>>>
>>> The level of stability on properties in general is unclear to me.
>>>
>>> Given that 115200 is standard for serial, it is unlikely to change
>>> though.. We can have an assert there instead?
>>>
>>> Peter, what do you think? thanks

IOW, until we merge Damien's "Clock framework API" series, I'd:

- rename 'baudbase' -> 'input_frequency_hz'

- set a 0 default value

  DEFINE_PROP_UINT32("input-frequency-hz", SerialState,
                      input_frequency_hz, 0),

- add a check in serial_realize()

     if (s->input_frequency_hz == 0) {
         error_setg(errp,
               "serial: input-frequency-hz property must be set");
         return;
     }

[*] https://www.mail-archive.com/qemu-devel@nongnu.org/msg642174.html

>> This property confused me by the past. It is _not_ the baudrate.
>> It is the input frequency clocking the UART ('XIN' pin, Xtal INput).
>>
>> Each board has its own frequency, and it can even be variable (the 
>> clock domain tree can reconfigure it at a different rate).
> 
> Laurent pointed me to the following commit which confirms my 
> interpretation:
> 
> $ git show 038eaf82c853
> commit 038eaf82c853f3bf8d4c106c0677bbf4adada7de
> Author: Stefan Weil <weil@mail.berlios.de>
> Date:   Sat Oct 31 11:28:11 2009 +0100
> 
>      serial: Add interface to set reference oscillator frequency
> 
>      Many (most?) serial interfaces have a programmable
>      clock which provides the reference frequency ("baudbase").
>      So a fixed baudbase which is only set once can be wrong.
> 
>      omap1.c is an example which could use the new interface
>      to change baudbase when the programmable clock changes.
>      ar7 system emulation (still not part of standard QEMU)
>      is similar to omap and already uses serial_set_frequency.
> 
>      Signed-off-by: Stefan Weil <weil@mail.berlios.de>
>      Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> 
> diff --git a/hw/pc.h b/hw/pc.h
> index 15fff8d103..03ffc91536 100644
> --- a/hw/pc.h
> +++ b/hw/pc.h
> @@ -13,6 +13,7 @@ SerialState *serial_mm_init (target_phys_addr_t base, 
> int it_shift,
>                                qemu_irq irq, int baudbase,
>                                CharDriverState *chr, int ioregister);
>   SerialState *serial_isa_init(int index, CharDriverState *chr);
> +void serial_set_frequency(SerialState *s, uint32_t frequency);
> 
>   /* parallel.c */
> 
> diff --git a/hw/serial.c b/hw/serial.c
> index fa12dcc075..0063260569 100644
> --- a/hw/serial.c
> +++ b/hw/serial.c
> @@ -730,6 +730,13 @@ static void serial_init_core(SerialState *s)
>                             serial_event, s);
>   }
> 
> +/* Change the main reference oscillator frequency. */
> +void serial_set_frequency(SerialState *s, uint32_t frequency)
> +{
> +    s->baudbase = frequency;
> +    serial_update_parameters(s);
> +}
> +

Re: [PATCH v4 17/37] mips: inline serial_init()

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 5568 bytes --]

On Monday, November 25, 2019, Marc-André Lureau <marcandre.lureau@redhat.com>
wrote:

> Hi
>
> On Mon, Nov 25, 2019 at 2:12 PM Aleksandar Markovic
> <aleksandar.m.mail@gmail.com> wrote:
> >
> >
> >
> > On Wednesday, November 20, 2019, Marc-André Lureau <
> marcandre.lureau@redhat.com> wrote:
> >>
> >> The function is specific to mipssim, let's inline it.
> >>
> >> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> >> ---
> >>  hw/char/serial.c         | 16 ----------------
> >>  hw/mips/mips_mipssim.c   | 15 ++++++++++++---
> >>  include/hw/char/serial.h |  2 --
> >>  3 files changed, 12 insertions(+), 21 deletions(-)
> >>
> >> diff --git a/hw/char/serial.c b/hw/char/serial.c
> >> index 164146ede8..23f0b02516 100644
> >> --- a/hw/char/serial.c
> >> +++ b/hw/char/serial.c
> >> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
> >>      .class_init = serial_io_class_init,
> >>  };
> >>
> >> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> >> -                         Chardev *chr, MemoryRegion *system_io)
> >> -{
> >> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> >> -
> >> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> >> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> >> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> >> -    qdev_init_nofail(DEVICE(sio));
> >> -
> >> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> >> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
> >> -
> >> -    return sio;
> >> -}
> >> -
> >>  static Property serial_properties[] = {
> >>      DEFINE_PROP_CHR("chardev", SerialState, chr),
> >>      DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
> >> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> >> index 282bbecb24..bfafa4d7e9 100644
> >> --- a/hw/mips/mips_mipssim.c
> >> +++ b/hw/mips/mips_mipssim.c
> >> @@ -40,6 +40,7 @@
> >>  #include "hw/loader.h"
> >>  #include "elf.h"
> >>  #include "hw/sysbus.h"
> >> +#include "hw/qdev-properties.h"
> >>  #include "exec/address-spaces.h"
> >>  #include "qemu/error-report.h"
> >>  #include "sysemu/qtest.h"
> >> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
> >>       * A single 16450 sits at offset 0x3f8. It is attached to
> >>       * MIPS CPU INT2, which is interrupt 4.
> >>       */
> >> -    if (serial_hd(0)) a
> >> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
> >> -                    get_system_io());
> >> +    if (serial_hd(0)) {
> >> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> >> +
> >> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> >> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> >> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> >> +        qdev_init_nofail(dev);
> >> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> >> +        memory_region_add_subregion(get_system_io(), 0x3f8,
> >> +                                    &SERIAL_IO(dev)->serial.io);
> >> +    }
> >
> >
> >
> > Please explain why the code in the deleted function and the new function
> are not identical. Why is the new code better?
>
> What differences do you spot?
>
> serial_init() looks very generic, but it is not, and this is
> confusing. QDev users should call qdev_create().
>
> If there is a single user, it is better to inline the code where it is
> being used, or move the function there.
>
>
Marc-Andre, you misunderstood me: I am not complaining about inlining at
all, I would do the same if I were you.

I just see you expressed the same thing from former function in a slightly
different way in the inlined version, and they look functionally equivalent
to me. But I suppose the new way is better, and I want to know why. Is it a
more standard way, or something even more? Please understand that I don't
come from device emulation background - I want to learn a little bit about
device-related QEMU customs and conventions, that is all. :)

For example,

in old version:

>> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
>> -                         Chardev *chr, MemoryRegion *system_io)
>> -{
>> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
>> -
>> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
>> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
>> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
>> -    qdev_init_nofail(DEVICE(sio));
>> -
>> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
>> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
>> -
>> -    return sio;
>> -}

in new version:

 >> +    if (serial_hd(0)) {
>> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>> +
>> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));   <-- you
didn't use here DEVICE(), while it is used in the previous line, and also
in the old version, and similar questions
>> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>> +        qdev_init_nofail(dev);
>> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
>> +        memory_region_add_subregion(get_system_io(), 0x3f8,
>> +                                    &SERIAL_IO(dev)->serial.io);
>> +    }

For you all these details are probably perfectly natural and obvois, but
not for me.


Yours, Aleksandar

[-- Attachment #2: Type: text/html, Size: 7687 bytes --]

Re: [PATCH v4 17/37] mips: inline serial_init()

[Marc-André Lureau]

Hi

On Mon, Nov 25, 2019 at 5:33 PM Aleksandar Markovic
<aleksandar.m.mail@gmail.com> wrote:
>
>
>
> On Monday, November 25, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
>>
>> Hi
>>
>> On Mon, Nov 25, 2019 at 2:12 PM Aleksandar Markovic
>> <aleksandar.m.mail@gmail.com> wrote:
>> >
>> >
>> >
>> > On Wednesday, November 20, 2019, Marc-André Lureau <marcandre.lureau@redhat.com> wrote:
>> >>
>> >> The function is specific to mipssim, let's inline it.
>> >>
>> >> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>> >> ---
>> >>  hw/char/serial.c         | 16 ----------------
>> >>  hw/mips/mips_mipssim.c   | 15 ++++++++++++---
>> >>  include/hw/char/serial.h |  2 --
>> >>  3 files changed, 12 insertions(+), 21 deletions(-)
>> >>
>> >> diff --git a/hw/char/serial.c b/hw/char/serial.c
>> >> index 164146ede8..23f0b02516 100644
>> >> --- a/hw/char/serial.c
>> >> +++ b/hw/char/serial.c
>> >> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
>> >>      .class_init = serial_io_class_init,
>> >>  };
>> >>
>> >> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
>> >> -                         Chardev *chr, MemoryRegion *system_io)
>> >> -{
>> >> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
>> >> -
>> >> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
>> >> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
>> >> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
>> >> -    qdev_init_nofail(DEVICE(sio));
>> >> -
>> >> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
>> >> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
>> >> -
>> >> -    return sio;
>> >> -}
>> >> -
>> >>  static Property serial_properties[] = {
>> >>      DEFINE_PROP_CHR("chardev", SerialState, chr),
>> >>      DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
>> >> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
>> >> index 282bbecb24..bfafa4d7e9 100644
>> >> --- a/hw/mips/mips_mipssim.c
>> >> +++ b/hw/mips/mips_mipssim.c
>> >> @@ -40,6 +40,7 @@
>> >>  #include "hw/loader.h"
>> >>  #include "elf.h"
>> >>  #include "hw/sysbus.h"
>> >> +#include "hw/qdev-properties.h"
>> >>  #include "exec/address-spaces.h"
>> >>  #include "qemu/error-report.h"
>> >>  #include "sysemu/qtest.h"
>> >> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
>> >>       * A single 16450 sits at offset 0x3f8. It is attached to
>> >>       * MIPS CPU INT2, which is interrupt 4.
>> >>       */
>> >> -    if (serial_hd(0)) a
>> >> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
>> >> -                    get_system_io());
>> >> +    if (serial_hd(0)) {
>> >> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
>> >> +
>> >> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
>> >> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
>> >> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>> >> +        qdev_init_nofail(dev);
>> >> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
>> >> +        memory_region_add_subregion(get_system_io(), 0x3f8,
>> >> +                                    &SERIAL_IO(dev)->serial.io);
>> >> +    }
>> >
>> >
>> >
>> > Please explain why the code in the deleted function and the new function are not identical. Why is the new code better?
>>
>> What differences do you spot?
>>
>> serial_init() looks very generic, but it is not, and this is
>> confusing. QDev users should call qdev_create().
>>
>> If there is a single user, it is better to inline the code where it is
>> being used, or move the function there.
>>
>
> Marc-Andre, you misunderstood me: I am not complaining about inlining at all, I would do the same if I were you.
>
> I just see you expressed the same thing from former function in a slightly different way in the inlined version, and they look functionally equivalent to me. But I suppose the new way is better, and I want to know why. Is it a more standard way, or something even more? Please understand that I don't come from device emulation background - I want to learn a little bit about device-related QEMU customs and conventions, that is all. :)
>
> For example,
>
> in old version:
>
> >> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> >> -                         Chardev *chr, MemoryRegion *system_io)
> >> -{
> >> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> >> -
> >> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> >> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> >> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> >> -    qdev_init_nofail(DEVICE(sio));
> >> -
> >> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> >> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
> >> -
> >> -    return sio;
> >> -}
>
> in new version:
>
>  >> +    if (serial_hd(0)) {
> >> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> >> +
> >> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> >> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));   <-- you didn't use here DEVICE(), while it is used in the previous line, and also in the old version, and similar questions
> >> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> >> +        qdev_init_nofail(dev);
> >> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> >> +        memory_region_add_subregion(get_system_io(), 0x3f8,
> >> +                                    &SERIAL_IO(dev)->serial.io);
> >> +    }
>
> For you all these details are probably perfectly natural and obvois, but not for me.

Oh, there is no big difference between a variable using the base type
DeviceState * or the instance type SerialIO *.

Basically, you need a variable to hold the instance, and then you cast
it with DEVICE() or SERIAL_IO() or event SYS_BUS_DEVICE() as
necessary, depending on what type the function you call expect.

I probably picked DeviceState* because it is the most commonly used,
and it avoids having plenty of DEVICE() casts.

Re: [PATCH v4 21/37] sm501: make SerialMM a child, export chardev property

[Marc-André Lureau]

On Thu, Nov 21, 2019 at 6:04 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Wed, 20 Nov 2019 at 15:31, Marc-André Lureau
> <marcandre.lureau@redhat.com> wrote:
> >
> > Embed the SerialMM sybus device, and re-export its "chardev" property.
> > That way, we can get rid of PROP_PTR "chr-state" and better track
> > devices relationship.
> >
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >  hw/display/sm501.c | 33 ++++++++++++++++++++++++---------
> >  hw/sh4/r2d.c       |  2 +-
> >  2 files changed, 25 insertions(+), 10 deletions(-)
> >
> > diff --git a/hw/display/sm501.c b/hw/display/sm501.c
> > index 1f33c87e65..c4445b28f9 100644
> > --- a/hw/display/sm501.c
> > +++ b/hw/display/sm501.c
> > @@ -1930,13 +1930,14 @@ typedef struct {
> >      SM501State state;
> >      uint32_t vram_size;
> >      uint32_t base;
> > -    void *chr_state;
> > +    SerialMM serial;
> >  } SM501SysBusState;
> >
> >  static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
> >  {
> >      SM501SysBusState *s = SYSBUS_SM501(dev);
> >      SysBusDevice *sbd = SYS_BUS_DEVICE(dev);
> > +    SerialState *ss = &s->serial.serial;
> >      DeviceState *usb_dev;
> >
> >      sm501_init(&s->state, dev, s->vram_size);
> > @@ -1958,17 +1959,19 @@ static void sm501_realize_sysbus(DeviceState *dev, Error **errp)
> >      sysbus_pass_irq(sbd, SYS_BUS_DEVICE(usb_dev));
> >
> >      /* bridge to serial emulation module */
> > -    if (s->chr_state) {
> > -        serial_mm_init(&s->state.mmio_region, SM501_UART0, 2,
> > -                       NULL, /* TODO : chain irq to IRL */
> > -                       115200, s->chr_state, DEVICE_LITTLE_ENDIAN);
> > +    /* FIXME: SM501_UART0 is always mapped, no need to check for the backend */
> > +    if (qemu_chr_fe_backend_connected(&ss->chr)) {
> > +        MemoryRegion *mr;
> > +        qdev_init_nofail(DEVICE(&s->serial));
> > +        mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->serial), 0);
> > +        memory_region_add_subregion(&s->state.mmio_region, SM501_UART0, mr);
> > +        /* TODO : chain irq to IRL */
> >      }
>
> I don't really understand what the FIXME is trying to
> tell me here. If we don't need to check for the backend,
> why is the code checking for it ? It means we have to fish
> around inside the SerialMM's implementation, which seems odd.
> Only mapping the UART registers if there happens to be a backend
> connected also doesn't conceptually seem like the right behaviour,
> because the registers should always exist. Since commit
> 12051d82f004024d5d the chardev mid-layer has correctly handled
> the backend not being connected (ie having a NULL chardev),
> so there's no longer any need for board/device code to special
> case the lack of a chardev.

As you noted, this is addressed by "[PATCH-for-5.0] hw/display/sm501:
Always map the UART0".

It's preferable to keep those 2 seperate. I can include Philippe
change on top, or earlier in my series, is that ok?

>
> >  }
> >
> >  static Property sm501_sysbus_properties[] = {
> >      DEFINE_PROP_UINT32("vram-size", SM501SysBusState, vram_size, 0),
> >      DEFINE_PROP_UINT32("base", SM501SysBusState, base, 0),
> > -    DEFINE_PROP_PTR("chr-state", SM501SysBusState, chr_state),
> >      DEFINE_PROP_END_OF_LIST(),
> >  };
> >
> > @@ -1999,9 +2002,20 @@ static void sm501_sysbus_class_init(ObjectClass *klass, void *data)
> >      dc->props = sm501_sysbus_properties;
> >      dc->reset = sm501_reset_sysbus;
> >      dc->vmsd = &vmstate_sm501_sysbus;
> > -    /* Note: pointer property "chr-state" may remain null, thus
> > -     * no need for dc->user_creatable = false;
> > -     */
> > +}
> > +
> > +static void sm501_sysbus_init(Object *o)
> > +{
> > +    SM501SysBusState *sm501 = SYSBUS_SM501(o);
> > +    SerialMM *smm = &sm501->serial;
> > +
> > +    sysbus_init_child_obj(o, "serial", smm, sizeof(SerialMM), TYPE_SERIAL_MM);
> > +    qdev_set_legacy_instance_id(DEVICE(smm), SM501_UART0, 2);
>
> The only board we use the sm501 sysbus device is the sh4 r2d
> board, and we don't care about migration compatibility there
> (indeed I would be unsurprised to find that it doesn't even work ;-))
> So I think we can reasonably not set the legacy-instance-ID
> and just declare in the commit message that this is a migration
> compat break for that board.
>

I am simply adapting serial_mm_init() code from "serial: register vmsd
with DeviceClass" here.

Dropping qdev_set_legacy_instance_id() is proabably a seperate concern.

> > +    qdev_prop_set_uint8(DEVICE(smm), "regshift", 2);
> > +    qdev_prop_set_uint8(DEVICE(smm), "endianness", DEVICE_LITTLE_ENDIAN);
> > +
> > +    object_property_add_alias(o, "chardev",
> > +                              OBJECT(smm), "chardev", &error_abort);
> >  }
> >
> >  static const TypeInfo sm501_sysbus_info = {
> > @@ -2009,6 +2023,7 @@ static const TypeInfo sm501_sysbus_info = {
> >      .parent        = TYPE_SYS_BUS_DEVICE,
> >      .instance_size = sizeof(SM501SysBusState),
> >      .class_init    = sm501_sysbus_class_init,
> > +    .instance_init = sm501_sysbus_init,
> >  };
> >
> >  #define TYPE_PCI_SM501 "sm501"
> > diff --git a/hw/sh4/r2d.c b/hw/sh4/r2d.c
> > index ee0840f380..72bb5285cc 100644
> > --- a/hw/sh4/r2d.c
> > +++ b/hw/sh4/r2d.c
> > @@ -272,7 +272,7 @@ static void r2d_init(MachineState *machine)
> >      busdev = SYS_BUS_DEVICE(dev);
> >      qdev_prop_set_uint32(dev, "vram-size", SM501_VRAM_SIZE);
> >      qdev_prop_set_uint32(dev, "base", 0x10000000);
> > -    qdev_prop_set_ptr(dev, "chr-state", serial_hd(2));
> > +    qdev_prop_set_chr(dev, "chardev", serial_hd(2));
> >      qdev_init_nofail(dev);
> >      sysbus_mmio_map(busdev, 0, 0x10000000);
> >      sysbus_mmio_map(busdev, 1, 0x13e00000);
> > --
> > 2.24.0
>
> thanks
> -- PMM
>


-- 
Marc-André Lureau

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Marc-André Lureau]

Hi

On Mon, Nov 25, 2019 at 5:04 PM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/25/19 1:54 PM, Philippe Mathieu-Daudé wrote:
> > On 11/25/19 12:26 PM, Philippe Mathieu-Daudé wrote:
> >> On 11/25/19 11:12 AM, Marc-André Lureau wrote:
> >>> Hi
> >>>
> >>> On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
> >>> <aleksandar.m.mail@gmail.com> wrote:
> >>>>
> >>>>
> >>>>
> >>>> On Wednesday, November 20, 2019, Marc-André Lureau
> >>>> <marcandre.lureau@redhat.com> wrote:
> >>>>>
> >>>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> >>>>> ---
> >>>>>   hw/mips/mips_mipssim.c | 1 -
> >>>>>   1 file changed, 1 deletion(-)
> >>>>>
> >>>>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> >>>>> index bfafa4d7e9..3cd0e6eb33 100644
> >>>>> --- a/hw/mips/mips_mipssim.c
> >>>>> +++ b/hw/mips/mips_mipssim.c
> >>>>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
> >>>>>       if (serial_hd(0)) {
> >>>>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> >>>>>
> >>>>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> >>>>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> >>>>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> >>>>>           qdev_init_nofail(dev);
> >>>>> --
> >>>>
> >>>>
> >>>> Please mention in your commit message where the default baudbase is
> >>>> set.
> >>>
> >>> ok
> >>>
> >>>> Also, is there a guarantie that default value 115200 will never
> >>>> change in future?
> >>>
> >>> The level of stability on properties in general is unclear to me.
> >>>
> >>> Given that 115200 is standard for serial, it is unlikely to change
> >>> though.. We can have an assert there instead?
> >>>
> >>> Peter, what do you think? thanks
>
> IOW, until we merge Damien's "Clock framework API" series, I'd:
>
> - rename 'baudbase' -> 'input_frequency_hz'
>
> - set a 0 default value
>
>   DEFINE_PROP_UINT32("input-frequency-hz", SerialState,
>                       input_frequency_hz, 0),
>
> - add a check in serial_realize()
>
>      if (s->input_frequency_hz == 0) {
>          error_setg(errp,
>                "serial: input-frequency-hz property must be set");
>          return;
>      }
>
> [*] https://www.mail-archive.com/qemu-devel@nongnu.org/msg642174.html
>

This is getting further away from this series goal, and my initial
goal. Let's add this to the backlog. I can drop a FIXME there.

> >> This property confused me by the past. It is _not_ the baudrate.
> >> It is the input frequency clocking the UART ('XIN' pin, Xtal INput).
> >>
> >> Each board has its own frequency, and it can even be variable (the
> >> clock domain tree can reconfigure it at a different rate).
> >
> > Laurent pointed me to the following commit which confirms my
> > interpretation:
> >
> > $ git show 038eaf82c853
> > commit 038eaf82c853f3bf8d4c106c0677bbf4adada7de
> > Author: Stefan Weil <weil@mail.berlios.de>
> > Date:   Sat Oct 31 11:28:11 2009 +0100
> >
> >      serial: Add interface to set reference oscillator frequency
> >
> >      Many (most?) serial interfaces have a programmable
> >      clock which provides the reference frequency ("baudbase").
> >      So a fixed baudbase which is only set once can be wrong.
> >
> >      omap1.c is an example which could use the new interface
> >      to change baudbase when the programmable clock changes.
> >      ar7 system emulation (still not part of standard QEMU)
> >      is similar to omap and already uses serial_set_frequency.
> >
> >      Signed-off-by: Stefan Weil <weil@mail.berlios.de>
> >      Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> >
> > diff --git a/hw/pc.h b/hw/pc.h
> > index 15fff8d103..03ffc91536 100644
> > --- a/hw/pc.h
> > +++ b/hw/pc.h
> > @@ -13,6 +13,7 @@ SerialState *serial_mm_init (target_phys_addr_t base,
> > int it_shift,
> >                                qemu_irq irq, int baudbase,
> >                                CharDriverState *chr, int ioregister);
> >   SerialState *serial_isa_init(int index, CharDriverState *chr);
> > +void serial_set_frequency(SerialState *s, uint32_t frequency);
> >
> >   /* parallel.c */
> >
> > diff --git a/hw/serial.c b/hw/serial.c
> > index fa12dcc075..0063260569 100644
> > --- a/hw/serial.c
> > +++ b/hw/serial.c
> > @@ -730,6 +730,13 @@ static void serial_init_core(SerialState *s)
> >                             serial_event, s);
> >   }
> >
> > +/* Change the main reference oscillator frequency. */
> > +void serial_set_frequency(SerialState *s, uint32_t frequency)
> > +{
> > +    s->baudbase = frequency;
> > +    serial_update_parameters(s);
> > +}
> > +
>

Re: [PATCH v4 18/37] mips: baudbase is 115200 by default

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 5320 bytes --]

On Wednesday, November 27, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> Hi
>
> On Mon, Nov 25, 2019 at 5:04 PM Philippe Mathieu-Daudé
> <philmd@redhat.com> wrote:
> >
> > On 11/25/19 1:54 PM, Philippe Mathieu-Daudé wrote:
> > > On 11/25/19 12:26 PM, Philippe Mathieu-Daudé wrote:
> > >> On 11/25/19 11:12 AM, Marc-André Lureau wrote:
> > >>> Hi
> > >>>
> > >>> On Mon, Nov 25, 2019 at 2:07 PM Aleksandar Markovic
> > >>> <aleksandar.m.mail@gmail.com> wrote:
> > >>>>
> > >>>>
> > >>>>
> > >>>> On Wednesday, November 20, 2019, Marc-André Lureau
> > >>>> <marcandre.lureau@redhat.com> wrote:
> > >>>>>
> > >>>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > >>>>> ---
> > >>>>>   hw/mips/mips_mipssim.c | 1 -
> > >>>>>   1 file changed, 1 deletion(-)
> > >>>>>
> > >>>>> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> > >>>>> index bfafa4d7e9..3cd0e6eb33 100644
> > >>>>> --- a/hw/mips/mips_mipssim.c
> > >>>>> +++ b/hw/mips/mips_mipssim.c
> > >>>>> @@ -223,7 +223,6 @@ mips_mipssim_init(MachineState *machine)
> > >>>>>       if (serial_hd(0)) {
> > >>>>>           DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> > >>>>>
> > >>>>> -        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> > >>>>>           qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> > >>>>>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> > >>>>>           qdev_init_nofail(dev);
> > >>>>> --
> > >>>>
> > >>>>
> > >>>> Please mention in your commit message where the default baudbase is
> > >>>> set.
> > >>>
> > >>> ok
> > >>>
> > >>>> Also, is there a guarantie that default value 115200 will never
> > >>>> change in future?
> > >>>
> > >>> The level of stability on properties in general is unclear to me.
> > >>>
> > >>> Given that 115200 is standard for serial, it is unlikely to change
> > >>> though.. We can have an assert there instead?
> > >>>
> > >>> Peter, what do you think? thanks
> >
> > IOW, until we merge Damien's "Clock framework API" series, I'd:
> >
> > - rename 'baudbase' -> 'input_frequency_hz'
> >
> > - set a 0 default value
> >
> >   DEFINE_PROP_UINT32("input-frequency-hz", SerialState,
> >                       input_frequency_hz, 0),
> >
> > - add a check in serial_realize()
> >
> >      if (s->input_frequency_hz == 0) {
> >          error_setg(errp,
> >                "serial: input-frequency-hz property must be set");
> >          return;
> >      }
> >
> > [*] https://www.mail-archive.com/qemu-devel@nongnu.org/msg642174.html
> >
>
> This is getting further away from this series goal, and my initial
> goal. Let's add this to the backlog. I can drop a FIXME there.


I agree. But please update commit message and/or add FIXME so that future
readers are given at least some background.

Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>


>
> > >> This property confused me by the past. It is _not_ the baudrate.
> > >> It is the input frequency clocking the UART ('XIN' pin, Xtal INput).
> > >>
> > >> Each board has its own frequency, and it can even be variable (the
> > >> clock domain tree can reconfigure it at a different rate).
> > >
> > > Laurent pointed me to the following commit which confirms my
> > > interpretation:
> > >
> > > $ git show 038eaf82c853
> > > commit 038eaf82c853f3bf8d4c106c0677bbf4adada7de
> > > Author: Stefan Weil <weil@mail.berlios.de>
> > > Date:   Sat Oct 31 11:28:11 2009 +0100
> > >
> > >      serial: Add interface to set reference oscillator frequency
> > >
> > >      Many (most?) serial interfaces have a programmable
> > >      clock which provides the reference frequency ("baudbase").
> > >      So a fixed baudbase which is only set once can be wrong.
> > >
> > >      omap1.c is an example which could use the new interface
> > >      to change baudbase when the programmable clock changes.
> > >      ar7 system emulation (still not part of standard QEMU)
> > >      is similar to omap and already uses serial_set_frequency.
> > >
> > >      Signed-off-by: Stefan Weil <weil@mail.berlios.de>
> > >      Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> > >
> > > diff --git a/hw/pc.h b/hw/pc.h
> > > index 15fff8d103..03ffc91536 100644
> > > --- a/hw/pc.h
> > > +++ b/hw/pc.h
> > > @@ -13,6 +13,7 @@ SerialState *serial_mm_init (target_phys_addr_t base,
> > > int it_shift,
> > >                                qemu_irq irq, int baudbase,
> > >                                CharDriverState *chr, int ioregister);
> > >   SerialState *serial_isa_init(int index, CharDriverState *chr);
> > > +void serial_set_frequency(SerialState *s, uint32_t frequency);
> > >
> > >   /* parallel.c */
> > >
> > > diff --git a/hw/serial.c b/hw/serial.c
> > > index fa12dcc075..0063260569 100644
> > > --- a/hw/serial.c
> > > +++ b/hw/serial.c
> > > @@ -730,6 +730,13 @@ static void serial_init_core(SerialState *s)
> > >                             serial_event, s);
> > >   }
> > >
> > > +/* Change the main reference oscillator frequency. */
> > > +void serial_set_frequency(SerialState *s, uint32_t frequency)
> > > +{
> > > +    s->baudbase = frequency;
> > > +    serial_update_parameters(s);
> > > +}
> > > +
> >
>
>

[-- Attachment #2: Type: text/html, Size: 7845 bytes --]

Re: [PATCH v4 17/37] mips: inline serial_init()

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 3586 bytes --]

On Wednesday, November 20, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> The function is specific to mipssim, let's inline it.
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/char/serial.c         | 16 ----------------
>  hw/mips/mips_mipssim.c   | 15 ++++++++++++---
>  include/hw/char/serial.h |  2 --
>  3 files changed, 12 insertions(+), 21 deletions(-)
>
>
Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>


> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 164146ede8..23f0b02516 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -1023,22 +1023,6 @@ static const TypeInfo serial_io_info = {
>      .class_init = serial_io_class_init,
>  };
>
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                         Chardev *chr, MemoryRegion *system_io)
> -{
> -    SerialIO *sio = SERIAL_IO(qdev_create(NULL, TYPE_SERIAL_IO));
> -
> -    qdev_prop_set_uint32(DEVICE(sio), "baudbase", baudbase);
> -    qdev_prop_set_chr(DEVICE(sio), "chardev", chr);
> -    qdev_set_legacy_instance_id(DEVICE(sio), base, 2);
> -    qdev_init_nofail(DEVICE(sio));
> -
> -    sysbus_connect_irq(SYS_BUS_DEVICE(sio), 0, irq);
> -    memory_region_add_subregion(system_io, base, &sio->serial.io);
> -
> -    return sio;
> -}
> -
>  static Property serial_properties[] = {
>      DEFINE_PROP_CHR("chardev", SerialState, chr),
>      DEFINE_PROP_UINT32("baudbase", SerialState, baudbase, 115200),
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 282bbecb24..bfafa4d7e9 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -40,6 +40,7 @@
>  #include "hw/loader.h"
>  #include "elf.h"
>  #include "hw/sysbus.h"
> +#include "hw/qdev-properties.h"
>  #include "exec/address-spaces.h"
>  #include "qemu/error-report.h"
>  #include "sysemu/qtest.h"
> @@ -219,9 +220,17 @@ mips_mipssim_init(MachineState *machine)
>       * A single 16450 sits at offset 0x3f8. It is attached to
>       * MIPS CPU INT2, which is interrupt 4.
>       */
> -    if (serial_hd(0))
> -        serial_init(0x3f8, env->irq[4], 115200, serial_hd(0),
> -                    get_system_io());
> +    if (serial_hd(0)) {
> +        DeviceState *dev = qdev_create(NULL, TYPE_SERIAL_IO);
> +
> +        qdev_prop_set_uint32(DEVICE(dev), "baudbase", 115200);
> +        qdev_prop_set_chr(dev, "chardev", serial_hd(0));
> +        qdev_set_legacy_instance_id(dev, 0x3f8, 2);
> +        qdev_init_nofail(dev);
> +        sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> +        memory_region_add_subregion(get_system_io(), 0x3f8,
> +                                    &SERIAL_IO(dev)->serial.io);
> +    }
>
>      if (nd_table[0].used)
>          /* MIPSnet uses the MIPS CPU INT0, which is interrupt 2. */
> diff --git a/include/hw/char/serial.h b/include/hw/char/serial.h
> index d356ba838c..535fa23a2b 100644
> --- a/include/hw/char/serial.h
> +++ b/include/hw/char/serial.h
> @@ -108,8 +108,6 @@ void serial_set_frequency(SerialState *s, uint32_t
> frequency);
>  #define TYPE_SERIAL_IO "serial-io"
>  #define SERIAL_IO(s) OBJECT_CHECK(SerialIO, (s), TYPE_SERIAL_IO)
>
> -SerialIO *serial_init(int base, qemu_irq irq, int baudbase,
> -                      Chardev *chr, MemoryRegion *system_io);
>  SerialMM *serial_mm_init(MemoryRegion *address_space,
>                           hwaddr base, int regshift,
>                           qemu_irq irq, int baudbase,
> --
> 2.24.0
>
>
>

[-- Attachment #2: Type: text/html, Size: 4676 bytes --]

Re: [PATCH v4 20/37] mips: use sysbus_mmio_get_region() instead of internal fields

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 1705 bytes --]

On Wednesday, November 20, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> Register the memory region with sysbus_init_mmio() and look it up with
> sysbus_mmio_get_region() to avoid accessing internal device fields.
>
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/char/serial.c       | 1 +
>  hw/mips/mips_mipssim.c | 3 ++-
>  2 files changed, 3 insertions(+), 1 deletion(-)
>
>
Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>


> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 23f0b02516..02c545ff8c 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -993,6 +993,7 @@ static void serial_io_realize(DeviceState *dev, Error
> **errp)
>      qdev_init_nofail(DEVICE(s));
>
>      memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
> +    sysbus_init_mmio(SYS_BUS_DEVICE(sio), &s->io);
>      sysbus_init_irq(SYS_BUS_DEVICE(sio), &s->irq);
>  }
>
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 2c2c7f25b2..84c03dd035 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -227,7 +227,8 @@ mips_mipssim_init(MachineState *machine)
>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>          qdev_init_nofail(dev);
>          sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> -        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->
> serial.io);
> +        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8,
> +                      sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0));
>      }
>
>      if (nd_table[0].used)
> --
> 2.24.0
>
>
>

[-- Attachment #2: Type: text/html, Size: 2453 bytes --]

Re: [PATCH v4 19/37] mips: use sysbus_add_io()

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 1208 bytes --]

On Wednesday, November 20, 2019, Marc-André Lureau <
marcandre.lureau@redhat.com> wrote:

> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/mips/mips_mipssim.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
>
I agree with the change, and with overall series in general, but please add
at least a sentence in the commit message here, explaining what is achieved
by the change.

With that sentence, certainly:

Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>


> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 3cd0e6eb33..2c2c7f25b2 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -227,8 +227,7 @@ mips_mipssim_init(MachineState *machine)
>          qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>          qdev_init_nofail(dev);
>          sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> -        memory_region_add_subregion(get_system_io(), 0x3f8,
> -                                    &SERIAL_IO(dev)->serial.io);
> +        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->
> serial.io);
>      }
>
>      if (nd_table[0].used)
> --
> 2.24.0
>
>
>

[-- Attachment #2: Type: text/html, Size: 1906 bytes --]

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Marc-André Lureau]

Hi Corey

On Fri, Nov 22, 2019 at 2:44 AM Corey Minyard <cminyard@mvista.com> wrote:
>
> On Wed, Nov 20, 2019 at 07:24:36PM +0400, Marc-André Lureau wrote:
> > Instead, set the initial data field directly.
> >
> > (the initial data is an array of 256 bytes. As I don't know if it may
> > change over time, I keep the pointer to original buffer as is, but it
> > might be worth to consider to copy it instead)
>
> This looks ok to me, I can add it to my tree.  I think we are in hard
> freeze now, so this will have to wait until that's done.

Thanks, I think I would rather keep it in this series, as the goal is
to remove PROP_PTR, and tracking various subsystems will be tedious.
Instead, we can hopefully do it in one go.

Does it get your reviewed-by then?

>
> -corey
>
> >
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >  hw/i2c/smbus_eeprom.c | 17 ++++++++---------
> >  1 file changed, 8 insertions(+), 9 deletions(-)
> >
> > diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
> > index 54c86a0112..533c728b3b 100644
> > --- a/hw/i2c/smbus_eeprom.c
> > +++ b/hw/i2c/smbus_eeprom.c
> > @@ -44,7 +44,7 @@
> >  typedef struct SMBusEEPROMDevice {
> >      SMBusDevice smbusdev;
> >      uint8_t data[SMBUS_EEPROM_SIZE];
> > -    void *init_data;
> > +    uint8_t *init_data;
> >      uint8_t offset;
> >      bool accessed;
> >  } SMBusEEPROMDevice;
> > @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
> >
> >  static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
> >  {
> > +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
> > +
> >      smbus_eeprom_reset(dev);
> > +    if (eeprom->init_data == NULL) {
> > +        error_setg(errp, "init_data cannot be NULL");
> > +    }
> >  }
> >
> > -static Property smbus_eeprom_properties[] = {
> > -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
> > -    DEFINE_PROP_END_OF_LIST(),
> > -};
> > -
> >  static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> >  {
> >      DeviceClass *dc = DEVICE_CLASS(klass);
> > @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> >      dc->reset = smbus_eeprom_reset;
> >      sc->receive_byte = eeprom_receive_byte;
> >      sc->write_data = eeprom_write_data;
> > -    dc->props = smbus_eeprom_properties;
> >      dc->vmsd = &vmstate_smbus_eeprom;
> > -    /* Reason: pointer property "data" */
> > +    /* Reason: init_data */
> >      dc->user_creatable = false;
> >  }
> >
> > @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
> >
> >      dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
> >      qdev_prop_set_uint8(dev, "address", address);
> > -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
> > +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
> >      qdev_init_nofail(dev);
> >  }
> >
> > --
> > 2.24.0
> >
>


-- 
Marc-André Lureau

Re: [PATCH v4 31/37] smbus-eeprom: remove PROP_PTR

[Corey Minyard]

On Fri, Nov 29, 2019 at 01:06:41PM +0400, Marc-André Lureau wrote:
> Hi Corey
> 
> On Fri, Nov 22, 2019 at 2:44 AM Corey Minyard <cminyard@mvista.com> wrote:
> >
> > On Wed, Nov 20, 2019 at 07:24:36PM +0400, Marc-André Lureau wrote:
> > > Instead, set the initial data field directly.
> > >
> > > (the initial data is an array of 256 bytes. As I don't know if it may
> > > change over time, I keep the pointer to original buffer as is, but it
> > > might be worth to consider to copy it instead)
> >
> > This looks ok to me, I can add it to my tree.  I think we are in hard
> > freeze now, so this will have to wait until that's done.
> 
> Thanks, I think I would rather keep it in this series, as the goal is
> to remove PROP_PTR, and tracking various subsystems will be tedious.
> Instead, we can hopefully do it in one go.
> 
> Does it get your reviewed-by then?

Yes, I thought I had done that.

Reviewed-by: Corey Minyard <cminyard@mvista.com>

> 
> >
> > -corey
> >
> > >
> > > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > > ---
> > >  hw/i2c/smbus_eeprom.c | 17 ++++++++---------
> > >  1 file changed, 8 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
> > > index 54c86a0112..533c728b3b 100644
> > > --- a/hw/i2c/smbus_eeprom.c
> > > +++ b/hw/i2c/smbus_eeprom.c
> > > @@ -44,7 +44,7 @@
> > >  typedef struct SMBusEEPROMDevice {
> > >      SMBusDevice smbusdev;
> > >      uint8_t data[SMBUS_EEPROM_SIZE];
> > > -    void *init_data;
> > > +    uint8_t *init_data;
> > >      uint8_t offset;
> > >      bool accessed;
> > >  } SMBusEEPROMDevice;
> > > @@ -129,14 +129,14 @@ static void smbus_eeprom_reset(DeviceState *dev)
> > >
> > >  static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
> > >  {
> > > +    SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
> > > +
> > >      smbus_eeprom_reset(dev);
> > > +    if (eeprom->init_data == NULL) {
> > > +        error_setg(errp, "init_data cannot be NULL");
> > > +    }
> > >  }
> > >
> > > -static Property smbus_eeprom_properties[] = {
> > > -    DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
> > > -    DEFINE_PROP_END_OF_LIST(),
> > > -};
> > > -
> > >  static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> > >  {
> > >      DeviceClass *dc = DEVICE_CLASS(klass);
> > > @@ -146,9 +146,8 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, void *data)
> > >      dc->reset = smbus_eeprom_reset;
> > >      sc->receive_byte = eeprom_receive_byte;
> > >      sc->write_data = eeprom_write_data;
> > > -    dc->props = smbus_eeprom_properties;
> > >      dc->vmsd = &vmstate_smbus_eeprom;
> > > -    /* Reason: pointer property "data" */
> > > +    /* Reason: init_data */
> > >      dc->user_creatable = false;
> > >  }
> > >
> > > @@ -172,7 +171,7 @@ void smbus_eeprom_init_one(I2CBus *smbus, uint8_t address, uint8_t *eeprom_buf)
> > >
> > >      dev = qdev_create((BusState *) smbus, TYPE_SMBUS_EEPROM);
> > >      qdev_prop_set_uint8(dev, "address", address);
> > > -    qdev_prop_set_ptr(dev, "data", eeprom_buf);
> > > +    SMBUS_EEPROM(dev)->init_data = eeprom_buf;
> > >      qdev_init_nofail(dev);
> > >  }
> > >
> > > --
> > > 2.24.0
> > >
> >
> 
> 
> -- 
> Marc-André Lureau

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

On Wed, Nov 20, 2019 at 7:25 PM Marc-André Lureau
<marcandre.lureau@redhat.com> wrote:
>
> Hi,
>
> QDEV_PROP_PTR is marked in multiple places as "FIXME/TODO/remove
> me". In most cases, it can be easily replaced with QDEV_PROP_LINK when
> the pointer points to an Object.
>
> There are a few places where such substitution isn't possible. For
> those places, it seems reasonable to use a specific setter method
> instead, and keep the user_creatable = false. In other places,
> proper usage of qdev or other facilies is the solution.
>
> The serial code wasn't converted to qdev, which makes it a bit more
> archaic to deal with. Let's convert it first, so we can more easily
> embed it from other devices, and re-export some properties and drop
> QDEV_PROP_PTR usage.

Before v5, is there any other comment for the following patches:

- "qdev: remove unused qdev_prop_int64"

I spotted dead-code, never used. Peter would rather keep it. Worth it?

- "chardev: generate an internal id when none given"

As explained, this is necessary for qdev_prop_set_chr()

- "serial: register vmsd with DeviceClass"

This is standard qdev-ification, however it breaks backward migration,
but that's just how qdev_set_legacy_instance_id() works.

- "RFC: mips/cps: fix setting saar property"

Perhaps I should have used FIX instead of RFC, because this should
actually be a real fix. However I could use someone help to exercise
the code path.

- "sm501: make SerialMM a child, export chardev property"

review?

- "qdev/qom: remove some TODO limitations now that PROP_PTR is gone"

This should be straightforward.

>
> v4: (after Peter & Philippe reviews)
>  - replaced "self" variable names with abbreviations
>  - split "mips: inline serial_init()"
>  - new patches: "mips: use sysbus_mmio_get_region() instead of internal
>    fields", "leon3: use qdev gpio facilities for the PIL", "qdev: use
>    g_strcmp0() instead of open-coding it", "qdev/qom: remove some TODO
>    limitations now that PROP_PTR is gone"
>  - dropped patches: "sparc: move PIL irq handling to cpu.c" & "serial:
>    add "instance-id" property"
>  - various comments / commit message tweaks
>  - added r-b tags
>
> v3:
>  - introduce SerialMM and SerialIO sysbus devices
>  - remove serial_mm_connect introduced in v2
>  - replace "base" property introduced in v2, use "instance-id" for
>    vmstate purpose only
>  - add a few preliminary clean-ups
>
> v2:
>  - qom-ify serial
>  - embed the serial from sm501, and expose a "chardev" property
>  - add "leon3: use qemu_irq framework instead of callback as property"
>  - add "sparc: move PIL irq handling to cpu.c"
>  - add "cris: improve passing PIC interrupt vector to the CPU"
>  - misc comment/todo changes, add r-b tags
>
> Marc-André Lureau (37):
>   qdev: remove unused qdev_prop_int64
>   sysbus: remove unused sysbus_try_create*
>   sysbus: remove outdated comment
>   chardev: generate an internal id when none given
>   serial-pci-multi: factor out multi_serial_get_port_count()
>   serial: initial qom-ification
>   serial: register vmsd with DeviceClass
>   serial: add "chardev" property
>   serial: add "baudbase" property
>   serial: realize the serial device
>   serial: replace serial_exit_core() with unrealize
>   serial: start making SerialMM a sysbus device
>   serial-mm: add "regshift" property
>   serial-mm: add endianness property
>   serial-mm: use sysbus facilities
>   serial: make SerialIO a sysbus device
>   mips: inline serial_init()
>   mips: baudbase is 115200 by default
>   mips: use sysbus_add_io()
>   mips: use sysbus_mmio_get_region() instead of internal fields
>   sm501: make SerialMM a child, export chardev property
>   vmmouse: replace PROP_PTR with PROP_LINK
>   lance: replace PROP_PTR with PROP_LINK
>   etraxfs: remove PROP_PTR usage
>   dp8393x: replace PROP_PTR with PROP_LINK
>   leon3: use qemu_irq framework instead of callback as property
>   leon3: use qdev gpio facilities for the PIL
>   qdev: use g_strcmp0() instead of open-coding it
>   RFC: mips/cps: fix setting saar property
>   cris: improve passing PIC interrupt vector to the CPU
>   smbus-eeprom: remove PROP_PTR
>   omap-intc: remove PROP_PTR
>   omap-i2c: remove PROP_PTR
>   omap-gpio: remove PROP_PTR
>   qdev: remove PROP_MEMORY_REGION
>   qdev: remove QDEV_PROP_PTR
>   qdev/qom: remove some TODO limitations now that PROP_PTR is gone
>
>  chardev/char.c               |  32 +++++--
>  hw/arm/omap1.c               |   8 +-
>  hw/arm/omap2.c               |  25 +++---
>  hw/char/omap_uart.c          |   2 +-
>  hw/char/serial-isa.c         |  12 ++-
>  hw/char/serial-pci-multi.c   |  55 +++++++-----
>  hw/char/serial-pci.c         |  18 +++-
>  hw/char/serial.c             | 170 ++++++++++++++++++++++++++++-------
>  hw/core/qdev-properties.c    |  50 -----------
>  hw/core/qdev.c               |  15 +---
>  hw/core/sysbus.c             |  32 -------
>  hw/cris/axis_dev88.c         |   4 -
>  hw/display/sm501.c           |  33 +++++--
>  hw/dma/sparc32_dma.c         |   2 +-
>  hw/gpio/omap_gpio.c          |  42 ++++-----
>  hw/i2c/omap_i2c.c            |  19 ++--
>  hw/i2c/smbus_eeprom.c        |  17 ++--
>  hw/i386/pc.c                 |   7 +-
>  hw/i386/vmmouse.c            |   8 +-
>  hw/input/pckbd.c             |   8 +-
>  hw/intc/etraxfs_pic.c        |  26 +-----
>  hw/intc/grlib_irqmp.c        |  35 +-------
>  hw/intc/omap_intc.c          |  17 ++--
>  hw/m68k/q800.c               |   3 +-
>  hw/mips/boston.c             |   2 +-
>  hw/mips/cps.c                |   2 +-
>  hw/mips/mips_jazz.c          |   3 +-
>  hw/mips/mips_malta.c         |   2 +-
>  hw/mips/mips_mipssim.c       |  14 ++-
>  hw/net/dp8393x.c             |   7 +-
>  hw/net/etraxfs_eth.c         |  44 ++++++---
>  hw/net/lance.c               |   5 +-
>  hw/net/pcnet-pci.c           |   2 +-
>  hw/net/pcnet.h               |   2 +-
>  hw/sh4/r2d.c                 |   2 +-
>  hw/sparc/leon3.c             |  15 +++-
>  include/hw/arm/omap.h        |  52 +++++++++++
>  include/hw/char/serial.h     |  43 ++++++---
>  include/hw/cris/etraxfs.h    |  20 +----
>  include/hw/input/i8042.h     |   4 +-
>  include/hw/qdev-properties.h |  27 ------
>  include/hw/sysbus.h          |  13 +--
>  include/qemu/id.h            |   1 +
>  qom/qom-qmp-cmds.c           |  10 ---
>  target/cris/cpu.c            |   8 ++
>  target/cris/cpu.h            |   1 +
>  util/id.c                    |   1 +
>  47 files changed, 497 insertions(+), 423 deletions(-)
>
> --
> 2.24.0
>
>


-- 
Marc-André Lureau

Re: [PATCH v6 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 886 bytes --]

On Sunday, December 1, 2019, Marc-André Lureau <marcandre.lureau@gmail.com>
wrote:


> - "RFC: mips/cps: fix setting saar property"
>
> Perhaps I should have used FIX instead of RFC, because this should
> actually be a real fix. However I could use someone help to exercise
> the code path.
>
>
Marc-André, hi.

There is a work in progress on fixing this. Can we in MIPS submit the fix
independently, since it involves some additional pieces of code that are
really deeply mips-specific? We acknowledge the bug, and want to develop
the real solution. Can you simply skip this RFC patch in your series, since
the issues will be handled separately in our patch, hopefully soon after
the merge window is open?

For all other mips parts of your series, you have my "reviewed-by"s , in
case I forgot to send them explicitely.

Regards, Aleksandar




> - "
>

[-- Attachment #2: Type: text/html, Size: 1352 bytes --]

Re: [PATCH v6 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi Aleksandar

On Sun, Dec 1, 2019 at 4:15 PM Aleksandar Markovic
<aleksandar.m.mail@gmail.com> wrote:
>
>
>
> On Sunday, December 1, 2019, Marc-André Lureau <marcandre.lureau@gmail.com> wrote:
>
>>
>> - "RFC: mips/cps: fix setting saar property"
>>
>> Perhaps I should have used FIX instead of RFC, because this should
>> actually be a real fix. However I could use someone help to exercise
>> the code path.
>>
>
> Marc-André, hi.
>
> There is a work in progress on fixing this. Can we in MIPS submit the fix independently, since it involves some additional pieces of code that are really deeply mips-specific? We acknowledge the bug, and want to develop the real solution. Can you simply skip this RFC patch in your series, since the issues will be handled separately in our patch, hopefully soon after the merge window is open?
>
> For all other mips parts of your series, you have my "reviewed-by"s , in case I forgot to send them explicitely.
>

This is a one-liner, and it is required to achieve the goal of the
series, to remove PROP_PTR.

If you prefer, I can instead comment the line with a FIXME, since it
is apparently broken anyway?

If you manage to get your fix merged earlier, then this patch can be
dropped. Else, is it a problem for the later fixes?

thanks


-- 
Marc-André Lureau

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Peter Maydell]

On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
> - "serial: register vmsd with DeviceClass"
>
> This is standard qdev-ification, however it breaks backward migration,
> but that's just how qdev_set_legacy_instance_id() works.

I don't understand this part. Surely the whole point
of setting a legacy instance ID is exactly to preserve
migration compatibility? If it doesn't do that then what
does setting legacy ID value do?

thanks
-- PMM

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

On Sun, Dec 1, 2019 at 9:18 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
> >
> > - "serial: register vmsd with DeviceClass"
> >
> > This is standard qdev-ification, however it breaks backward migration,
> > but that's just how qdev_set_legacy_instance_id() works.
>
> I don't understand this part. Surely the whole point
> of setting a legacy instance ID is exactly to preserve
> migration compatibility? If it doesn't do that then what
> does setting legacy ID value do?
>

It works in old->new direction only, because new code can match the
legacy instance id.

But when going from new->old, the legacy instance id is lost, as it
uses new 0-based instance_id.

This is just the way we have done so far, but I wish to be corrected
if I am wrong.

See the commit description for a bit more details.


-- 
Marc-André Lureau

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Peter Maydell]

On Sun, 1 Dec 2019 at 17:27, Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
> Hi
>
> On Sun, Dec 1, 2019 at 9:18 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> >
> > On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
> > <marcandre.lureau@gmail.com> wrote:
> > >
> > > - "serial: register vmsd with DeviceClass"
> > >
> > > This is standard qdev-ification, however it breaks backward migration,
> > > but that's just how qdev_set_legacy_instance_id() works.
> >
> > I don't understand this part. Surely the whole point
> > of setting a legacy instance ID is exactly to preserve
> > migration compatibility? If it doesn't do that then what
> > does setting legacy ID value do?
> >
>
> It works in old->new direction only, because new code can match the
> legacy instance id.
>
> But when going from new->old, the legacy instance id is lost, as it
> uses new 0-based instance_id.

I still don't understand. My mental model of the situation is:

 * in the old (current) version of the code, the instance ID
   is some random thing resulting from what the old code does
 * in the new version of the code, we use qdev_set_legacy_instance_id,
   and so instead of using the ID you'd naturally get as a
   written-from-scratch qdev device, it uses the legacy value
   you pass in
 * thus the device/board in both old and new versions of QEMU
   uses the same value and migration in both directions works

I don't understand why we would ever be using a "new 0-based
instance_id" -- it seems to me that the whole point of setting
a legacy ID value is that we will use it always, and I don't
understand how the board code can know that it's going to be
the target of an old->new migration as opposed to being the
source of a new->old migration such that it can end up with
a different ID value in the latter case.

If qdev_set_legacy_instance_id() doesn't work the way I
think it does above, what *does* it do ?

thanks
-- PMM

Re: [PATCH v6 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 1823 bytes --]

On Sunday, December 1, 2019, Marc-André Lureau <marcandre.lureau@gmail.com>
wrote:

> Hi Aleksandar
>
> On Sun, Dec 1, 2019 at 4:15 PM Aleksandar Markovic
> <aleksandar.m.mail@gmail.com> wrote:
> >
> >
> >
> > On Sunday, December 1, 2019, Marc-André Lureau <
> marcandre.lureau@gmail.com> wrote:
> >
> >>
> >> - "RFC: mips/cps: fix setting saar property"
> >>
> >> Perhaps I should have used FIX instead of RFC, because this should
> >> actually be a real fix. However I could use someone help to exercise
> >> the code path.
> >>
> >
> > Marc-André, hi.
> >
> > There is a work in progress on fixing this. Can we in MIPS submit the
> fix independently, since it involves some additional pieces of code that
> are really deeply mips-specific? We acknowledge the bug, and want to
> develop the real solution. Can you simply skip this RFC patch in your
> series, since the issues will be handled separately in our patch, hopefully
> soon after the merge window is open?
> >
> > For all other mips parts of your series, you have my "reviewed-by"s , in
> case I forgot to send them explicitely.
> >
>
> This is a one-liner, and it is required to achieve the goal of the
> series, to remove PROP_PTR.
>
> If you prefer, I can instead comment the line with a FIXME, since it
> is apparently broken anyway?
>
> If you manage to get your fix merged earlier, then this patch can be
> dropped. Else, is it a problem for the later fixes?
>
>
OK, Marc-André,

Please go ahead with this patch, so that the goal of the series is
achieved, and we will later submitt a wider patch that will address the
root cause. Just remove RFC from subject, everything else looks fine to me.
You can add my "reviewed-by".

Yours, Aleksandar




> thanks
>
>
> --
> Marc-André Lureau
>

[-- Attachment #2: Type: text/html, Size: 2489 bytes --]

Re: [PATCH v6 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Aleksandar Markovic]

[-- Attachment #1: Type: text/plain, Size: 2245 bytes --]

On Sunday, December 1, 2019, Aleksandar Markovic <
aleksandar.m.mail@gmail.com> wrote:

>
>
> On Sunday, December 1, 2019, Marc-André Lureau <marcandre.lureau@gmail.com>
> wrote:
>
>> Hi Aleksandar
>>
>> On Sun, Dec 1, 2019 at 4:15 PM Aleksandar Markovic
>> <aleksandar.m.mail@gmail.com> wrote:
>> >
>> >
>> >
>> > On Sunday, December 1, 2019, Marc-André Lureau <
>> marcandre.lureau@gmail.com> wrote:
>> >
>> >>
>> >> - "RFC: mips/cps: fix setting saar property"
>> >>
>> >> Perhaps I should have used FIX instead of RFC, because this should
>> >> actually be a real fix. However I could use someone help to exercise
>> >> the code path.
>> >>
>> >
>> > Marc-André, hi.
>> >
>> > There is a work in progress on fixing this. Can we in MIPS submit the
>> fix independently, since it involves some additional pieces of code that
>> are really deeply mips-specific? We acknowledge the bug, and want to
>> develop the real solution. Can you simply skip this RFC patch in your
>> series, since the issues will be handled separately in our patch, hopefully
>> soon after the merge window is open?
>> >
>> > For all other mips parts of your series, you have my "reviewed-by"s ,
>> in case I forgot to send them explicitely.
>> >
>>
>> This is a one-liner, and it is required to achieve the goal of the
>> series, to remove PROP_PTR.
>>
>> If you prefer, I can instead comment the line with a FIXME, since it
>> is apparently broken anyway?
>>
>> If you manage to get your fix merged earlier, then this patch can be
>> dropped. Else, is it a problem for the later fixes?
>>
>>
> OK, Marc-André,
>
> Please go ahead with this patch, so that the goal of the series is
> achieved, and we will later submitt a wider patch that will address the
> root cause. Just remove RFC from subject, everything else looks fine to me.
> You can add my "reviewed-by".
>
>
I mean, yes, you are right, it is broken, with or without the patch, so go
ahead, at least your series will fulfill its purpose, and I'll have enough
time to integrate the fix later on, without interfering each other.

Thanks for the series!


Yours, Aleksandar
>
>
>
>
>> thanks
>>
>>
>> --
>> Marc-André Lureau
>>
>

[-- Attachment #2: Type: text/html, Size: 3246 bytes --]

Re: [PATCH v4 01/37] qdev: remove unused qdev_prop_int64

[Markus Armbruster]

Marc-André Lureau <marcandre.lureau@redhat.com> writes:

> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/core/qdev-properties.c    | 32 --------------------------------
>  include/hw/qdev-properties.h |  3 ---
>  2 files changed, 35 deletions(-)
>
> diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c
> index ac28890e5a..be4cb01f0b 100644
> --- a/hw/core/qdev-properties.c
> +++ b/hw/core/qdev-properties.c
> @@ -409,31 +409,6 @@ static void set_uint64(Object *obj, Visitor *v, const char *name,
>      visit_type_uint64(v, name, ptr, errp);
>  }
>  
> -static void get_int64(Object *obj, Visitor *v, const char *name,
> -                      void *opaque, Error **errp)
> -{
> -    DeviceState *dev = DEVICE(obj);
> -    Property *prop = opaque;
> -    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
> -
> -    visit_type_int64(v, name, ptr, errp);
> -}
> -
> -static void set_int64(Object *obj, Visitor *v, const char *name,
> -                      void *opaque, Error **errp)
> -{
> -    DeviceState *dev = DEVICE(obj);
> -    Property *prop = opaque;
> -    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
> -
> -    if (dev->realized) {
> -        qdev_prop_set_after_realize(dev, name, errp);
> -        return;
> -    }
> -
> -    visit_type_int64(v, name, ptr, errp);
> -}
> -
>  const PropertyInfo qdev_prop_uint64 = {
>      .name  = "uint64",
>      .get   = get_uint64,
> @@ -441,13 +416,6 @@ const PropertyInfo qdev_prop_uint64 = {
>      .set_default_value = set_default_value_uint,
>  };
>  
> -const PropertyInfo qdev_prop_int64 = {
> -    .name  = "int64",
> -    .get   = get_int64,
> -    .set   = set_int64,
> -    .set_default_value = set_default_value_int,
> -};
> -
>  /* --- string --- */
>  
>  static void release_string(Object *obj, const char *name, void *opaque)
> diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h
> index c6a8cb5516..690ff07ae2 100644
> --- a/include/hw/qdev-properties.h
> +++ b/include/hw/qdev-properties.h
> @@ -13,7 +13,6 @@ extern const PropertyInfo qdev_prop_uint16;
>  extern const PropertyInfo qdev_prop_uint32;
>  extern const PropertyInfo qdev_prop_int32;
>  extern const PropertyInfo qdev_prop_uint64;
> -extern const PropertyInfo qdev_prop_int64;
>  extern const PropertyInfo qdev_prop_size;
>  extern const PropertyInfo qdev_prop_string;
>  extern const PropertyInfo qdev_prop_chr;
> @@ -164,8 +163,6 @@ extern const PropertyInfo qdev_prop_pcie_link_width;
>      DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int32, int32_t)
>  #define DEFINE_PROP_UINT64(_n, _s, _f, _d)                      \
>      DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_uint64, uint64_t)
> -#define DEFINE_PROP_INT64(_n, _s, _f, _d)                      \
> -    DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int64, int64_t)
>  #define DEFINE_PROP_SIZE(_n, _s, _f, _d)                       \
>      DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_size, uint64_t)
>  #define DEFINE_PROP_PCI_DEVFN(_n, _s, _f, _d)                   \

History of its use:

    Author: Peter Xu <peterx@redhat.com>
    Date:   Tue Jul 18 11:39:01 2017 +0800

        qdev: provide DEFINE_PROP_INT64()

        We have nearly all the stuff, but this one is missing. Add it in.

        Am going to use this new helper for MigrationParameters fields, since
        most of them are int64_t.

        CC: Markus Armbruster <armbru@redhat.com>
        CC: Eduardo Habkost <ehabkost@redhat.com>
        CC: Marc-André Lureau <marcandre.lureau@redhat.com>
        CC: Peter Xu <peterx@redhat.com>
        CC: Juan Quintela <quintela@redhat.com>
        CC: Marcel Apfelbaum <marcel@redhat.com>
        CC: Eric Blake <eblake@redhat.com>
        Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
        Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
        Reviewed-by: Juan Quintela <quintela@redhat.com>
        Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
        Signed-off-by: Peter Xu <peterx@redhat.com>
        Message-Id: <1500349150-13240-2-git-send-email-peterx@redhat.com>
        Signed-off-by: Juan Quintela <quintela@redhat.com>

    commit 89632fafdc64927647b6f45416307cd3d4c746db
    Author: Peter Xu <peterx@redhat.com>
    Date:   Tue Jul 18 11:39:02 2017 +0800

        migration: export parameters to props

        Export migration parameters to qdev properties. Then we can use, for
        example:

          -global migration.x-cpu-throttle-initial=xxx

        To specify migration parameters during init.

        Prefix "x-" is appended for each parameter exported to show that this is
        not a stable interface, and only for debugging/testing purpose.

        Reviewed-by: Juan Quintela <quintela@redhat.com>
        Signed-off-by: Peter Xu <peterx@redhat.com>
        Message-Id: <1500349150-13240-3-git-send-email-peterx@redhat.com>
        Signed-off-by: Juan Quintela <quintela@redhat.com>

    [More commits adding uses...]

    commit 741d4086c856320807a2575389d7c0505578270b
    Author: Juan Quintela <quintela@redhat.com>
    Date:   Fri Dec 1 13:08:38 2017 +0100

        migration: Use proper types in json

        We use int for everything (int64_t), and then we check that value is
        between 0 and 255.  Change it to the valid types.

        This change only happens for HMP.  QMP always use bytes and similar.

        Signed-off-by: Juan Quintela <quintela@redhat.com>
        Reviewed-by: Eric Blake <eblake@redhat.com>

    [All uses gone again]

We can remove it now, and revert the removal when new uses pop up.
Churn.  I'd say remove if we think new uses are unlikely.  I doubt they
are.

Re: [PATCH v4 01/37] qdev: remove unused qdev_prop_int64

[Marc-André Lureau]

Hi

On Mon, Dec 2, 2019 at 9:23 AM Markus Armbruster <armbru@redhat.com> wrote:
>
> Marc-André Lureau <marcandre.lureau@redhat.com> writes:
>
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >  hw/core/qdev-properties.c    | 32 --------------------------------
> >  include/hw/qdev-properties.h |  3 ---
> >  2 files changed, 35 deletions(-)
> >
> > diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c
> > index ac28890e5a..be4cb01f0b 100644
> > --- a/hw/core/qdev-properties.c
> > +++ b/hw/core/qdev-properties.c
> > @@ -409,31 +409,6 @@ static void set_uint64(Object *obj, Visitor *v, const char *name,
> >      visit_type_uint64(v, name, ptr, errp);
> >  }
> >
> > -static void get_int64(Object *obj, Visitor *v, const char *name,
> > -                      void *opaque, Error **errp)
> > -{
> > -    DeviceState *dev = DEVICE(obj);
> > -    Property *prop = opaque;
> > -    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
> > -
> > -    visit_type_int64(v, name, ptr, errp);
> > -}
> > -
> > -static void set_int64(Object *obj, Visitor *v, const char *name,
> > -                      void *opaque, Error **errp)
> > -{
> > -    DeviceState *dev = DEVICE(obj);
> > -    Property *prop = opaque;
> > -    int64_t *ptr = qdev_get_prop_ptr(dev, prop);
> > -
> > -    if (dev->realized) {
> > -        qdev_prop_set_after_realize(dev, name, errp);
> > -        return;
> > -    }
> > -
> > -    visit_type_int64(v, name, ptr, errp);
> > -}
> > -
> >  const PropertyInfo qdev_prop_uint64 = {
> >      .name  = "uint64",
> >      .get   = get_uint64,
> > @@ -441,13 +416,6 @@ const PropertyInfo qdev_prop_uint64 = {
> >      .set_default_value = set_default_value_uint,
> >  };
> >
> > -const PropertyInfo qdev_prop_int64 = {
> > -    .name  = "int64",
> > -    .get   = get_int64,
> > -    .set   = set_int64,
> > -    .set_default_value = set_default_value_int,
> > -};
> > -
> >  /* --- string --- */
> >
> >  static void release_string(Object *obj, const char *name, void *opaque)
> > diff --git a/include/hw/qdev-properties.h b/include/hw/qdev-properties.h
> > index c6a8cb5516..690ff07ae2 100644
> > --- a/include/hw/qdev-properties.h
> > +++ b/include/hw/qdev-properties.h
> > @@ -13,7 +13,6 @@ extern const PropertyInfo qdev_prop_uint16;
> >  extern const PropertyInfo qdev_prop_uint32;
> >  extern const PropertyInfo qdev_prop_int32;
> >  extern const PropertyInfo qdev_prop_uint64;
> > -extern const PropertyInfo qdev_prop_int64;
> >  extern const PropertyInfo qdev_prop_size;
> >  extern const PropertyInfo qdev_prop_string;
> >  extern const PropertyInfo qdev_prop_chr;
> > @@ -164,8 +163,6 @@ extern const PropertyInfo qdev_prop_pcie_link_width;
> >      DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int32, int32_t)
> >  #define DEFINE_PROP_UINT64(_n, _s, _f, _d)                      \
> >      DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_uint64, uint64_t)
> > -#define DEFINE_PROP_INT64(_n, _s, _f, _d)                      \
> > -    DEFINE_PROP_SIGNED(_n, _s, _f, _d, qdev_prop_int64, int64_t)
> >  #define DEFINE_PROP_SIZE(_n, _s, _f, _d)                       \
> >      DEFINE_PROP_UNSIGNED(_n, _s, _f, _d, qdev_prop_size, uint64_t)
> >  #define DEFINE_PROP_PCI_DEVFN(_n, _s, _f, _d)                   \
>
> History of its use:
>
>     Author: Peter Xu <peterx@redhat.com>
>     Date:   Tue Jul 18 11:39:01 2017 +0800
>
>         qdev: provide DEFINE_PROP_INT64()
>
>         We have nearly all the stuff, but this one is missing. Add it in.
>
>         Am going to use this new helper for MigrationParameters fields, since
>         most of them are int64_t.
>
>         CC: Markus Armbruster <armbru@redhat.com>
>         CC: Eduardo Habkost <ehabkost@redhat.com>
>         CC: Marc-André Lureau <marcandre.lureau@redhat.com>
>         CC: Peter Xu <peterx@redhat.com>
>         CC: Juan Quintela <quintela@redhat.com>
>         CC: Marcel Apfelbaum <marcel@redhat.com>
>         CC: Eric Blake <eblake@redhat.com>
>         Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>         Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
>         Reviewed-by: Juan Quintela <quintela@redhat.com>
>         Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
>         Signed-off-by: Peter Xu <peterx@redhat.com>
>         Message-Id: <1500349150-13240-2-git-send-email-peterx@redhat.com>
>         Signed-off-by: Juan Quintela <quintela@redhat.com>
>
>     commit 89632fafdc64927647b6f45416307cd3d4c746db
>     Author: Peter Xu <peterx@redhat.com>
>     Date:   Tue Jul 18 11:39:02 2017 +0800
>
>         migration: export parameters to props
>
>         Export migration parameters to qdev properties. Then we can use, for
>         example:
>
>           -global migration.x-cpu-throttle-initial=xxx
>
>         To specify migration parameters during init.
>
>         Prefix "x-" is appended for each parameter exported to show that this is
>         not a stable interface, and only for debugging/testing purpose.
>
>         Reviewed-by: Juan Quintela <quintela@redhat.com>
>         Signed-off-by: Peter Xu <peterx@redhat.com>
>         Message-Id: <1500349150-13240-3-git-send-email-peterx@redhat.com>
>         Signed-off-by: Juan Quintela <quintela@redhat.com>
>
>     [More commits adding uses...]
>
>     commit 741d4086c856320807a2575389d7c0505578270b
>     Author: Juan Quintela <quintela@redhat.com>
>     Date:   Fri Dec 1 13:08:38 2017 +0100
>
>         migration: Use proper types in json
>
>         We use int for everything (int64_t), and then we check that value is
>         between 0 and 255.  Change it to the valid types.
>
>         This change only happens for HMP.  QMP always use bytes and similar.
>
>         Signed-off-by: Juan Quintela <quintela@redhat.com>
>         Reviewed-by: Eric Blake <eblake@redhat.com>
>
>     [All uses gone again]
>
> We can remove it now, and revert the removal when new uses pop up.
> Churn.  I'd say remove if we think new uses are unlikely.  I doubt they
> are.
>

Oh, somehow I missed those users. I thought the code had never been
exercised. Ok, let's keep it for now.

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

On Sun, Dec 1, 2019 at 10:10 PM Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Sun, 1 Dec 2019 at 17:27, Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
> >
> > Hi
> >
> > On Sun, Dec 1, 2019 at 9:18 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > >
> > > On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
> > > <marcandre.lureau@gmail.com> wrote:
> > > >
> > > > - "serial: register vmsd with DeviceClass"
> > > >
> > > > This is standard qdev-ification, however it breaks backward migration,
> > > > but that's just how qdev_set_legacy_instance_id() works.
> > >
> > > I don't understand this part. Surely the whole point
> > > of setting a legacy instance ID is exactly to preserve
> > > migration compatibility? If it doesn't do that then what
> > > does setting legacy ID value do?
> > >
> >
> > It works in old->new direction only, because new code can match the
> > legacy instance id.
> >
> > But when going from new->old, the legacy instance id is lost, as it
> > uses new 0-based instance_id.
>
> I still don't understand. My mental model of the situation is:
>
>  * in the old (current) version of the code, the instance ID
>    is some random thing resulting from what the old code does

right

>  * in the new version of the code, we use qdev_set_legacy_instance_id,
>    and so instead of using the ID you'd naturally get as a
>    written-from-scratch qdev device, it uses the legacy value
>    you pass in

no, it only sets the SaveStateEntry.alias_id, which is only used
during incoming migration in find_se().

Iow, it only works old->new.

>  * thus the device/board in both old and new versions of QEMU
>    uses the same value and migration in both directions works

sadly no

>
> I don't understand why we would ever be using a "new 0-based
> instance_id" -- it seems to me that the whole point of setting
> a legacy ID value is that we will use it always, and I don't
> understand how the board code can know that it's going to be
> the target of an old->new migration as opposed to being the
> source of a new->old migration such that it can end up with
> a different ID value in the latter case.

The target will find the "legacy" alias with find_se() on incoming
migration, but any new outgoing migration will use the new 0-based
instance_id

>
> If qdev_set_legacy_instance_id() doesn't work the way I
> think it does above, what *does* it do ?

just set the old alias_id for incoming migration.

David, is that correct?

thanks


-- 
Marc-André Lureau

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

On Sun, Dec 1, 2019 at 2:19 PM Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
>
> - "chardev: generate an internal id when none given"
>
> As explained, this is necessary for qdev_prop_set_chr()

ping

>
> - "serial: register vmsd with DeviceClass"
>
> This is standard qdev-ification, however it breaks backward migration,
> but that's just how qdev_set_legacy_instance_id() works.

See thread, someone could review or nack (if backward migration is a problem).

>
> - "sm501: make SerialMM a child, export chardev property"
>
> review?

ping

>
> - "qdev/qom: remove some TODO limitations now that PROP_PTR is gone"
>
> This should be straightforward.

ping

thanks

-- 
Marc-André Lureau

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Dr. David Alan Gilbert]

Apologies for the delay.

* Marc-André Lureau (marcandre.lureau@gmail.com) wrote:
> Hi
> 
> On Sun, Dec 1, 2019 at 10:10 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> >
> > On Sun, 1 Dec 2019 at 17:27, Marc-André Lureau
> > <marcandre.lureau@gmail.com> wrote:
> > >
> > > Hi
> > >
> > > On Sun, Dec 1, 2019 at 9:18 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > > >
> > > > On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
> > > > <marcandre.lureau@gmail.com> wrote:
> > > > >
> > > > > - "serial: register vmsd with DeviceClass"
> > > > >
> > > > > This is standard qdev-ification, however it breaks backward migration,
> > > > > but that's just how qdev_set_legacy_instance_id() works.
> > > >
> > > > I don't understand this part. Surely the whole point
> > > > of setting a legacy instance ID is exactly to preserve
> > > > migration compatibility? If it doesn't do that then what
> > > > does setting legacy ID value do?
> > > >
> > >
> > > It works in old->new direction only, because new code can match the
> > > legacy instance id.
> > >
> > > But when going from new->old, the legacy instance id is lost, as it
> > > uses new 0-based instance_id.
> >
> > I still don't understand. My mental model of the situation is:
> >
> >  * in the old (current) version of the code, the instance ID
> >    is some random thing resulting from what the old code does
> 
> right
> 
> >  * in the new version of the code, we use qdev_set_legacy_instance_id,
> >    and so instead of using the ID you'd naturally get as a
> >    written-from-scratch qdev device, it uses the legacy value
> >    you pass in
> 
> no, it only sets the SaveStateEntry.alias_id, which is only used
> during incoming migration in find_se().
> 
> Iow, it only works old->new.
> 
> >  * thus the device/board in both old and new versions of QEMU
> >    uses the same value and migration in both directions works
> 
> sadly no
> 
> >
> > I don't understand why we would ever be using a "new 0-based
> > instance_id" -- it seems to me that the whole point of setting
> > a legacy ID value is that we will use it always, and I don't
> > understand how the board code can know that it's going to be
> > the target of an old->new migration as opposed to being the
> > source of a new->old migration such that it can end up with
> > a different ID value in the latter case.
> 
> The target will find the "legacy" alias with find_se() on incoming
> migration, but any new outgoing migration will use the new 0-based
> instance_id
> 
> >
> > If qdev_set_legacy_instance_id() doesn't work the way I
> > think it does above, what *does* it do ?
> 
> just set the old alias_id for incoming migration.
> 
> David, is that correct?

Yes, I think it is.
However, I'm curious which devices you're finding are explicitly setting
their id's;  there aren't many - although there are some that probably
should!
For example, running an x86 image with:
   -device isa-parallel,chardev=... -device isa-serial -device isa-serial -trace enable=qemu_loadvm_state_section_startfull

shows:
qemu_loadvm_state_section_startfull(uint32_t section_id, const char *idstr, uint32_t instance_id, uint32_t version_id) "%u(%s) %u %u"

165217@1576179638.856300:qemu_loadvm_state_section_startfull 41(serial) 0 3
165217@1576179638.856307:qemu_loadvm_state_section_startfull 42(serial) 1 3
165217@1576179638.856311:qemu_loadvm_state_section_startfull 43(parallel_isa) 0 1

so those two serial devices are instances '0' and '1' I think by luck of
their command line order, rather than having specified their base
address (which would have been safer).

Dave



> thanks
> 
> 
> -- 
> Marc-André Lureau
> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

On Fri, Dec 13, 2019 at 12:18 AM Dr. David Alan Gilbert
<dgilbert@redhat.com> wrote:
>
> Apologies for the delay.
>
> * Marc-André Lureau (marcandre.lureau@gmail.com) wrote:
> > Hi
> >
> > On Sun, Dec 1, 2019 at 10:10 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > >
> > > On Sun, 1 Dec 2019 at 17:27, Marc-André Lureau
> > > <marcandre.lureau@gmail.com> wrote:
> > > >
> > > > Hi
> > > >
> > > > On Sun, Dec 1, 2019 at 9:18 PM Peter Maydell <peter.maydell@linaro.org> wrote:
> > > > >
> > > > > On Sun, 1 Dec 2019 at 10:19, Marc-André Lureau
> > > > > <marcandre.lureau@gmail.com> wrote:
> > > > > >
> > > > > > - "serial: register vmsd with DeviceClass"
> > > > > >
> > > > > > This is standard qdev-ification, however it breaks backward migration,
> > > > > > but that's just how qdev_set_legacy_instance_id() works.
> > > > >
> > > > > I don't understand this part. Surely the whole point
> > > > > of setting a legacy instance ID is exactly to preserve
> > > > > migration compatibility? If it doesn't do that then what
> > > > > does setting legacy ID value do?
> > > > >
> > > >
> > > > It works in old->new direction only, because new code can match the
> > > > legacy instance id.
> > > >
> > > > But when going from new->old, the legacy instance id is lost, as it
> > > > uses new 0-based instance_id.
> > >
> > > I still don't understand. My mental model of the situation is:
> > >
> > >  * in the old (current) version of the code, the instance ID
> > >    is some random thing resulting from what the old code does
> >
> > right
> >
> > >  * in the new version of the code, we use qdev_set_legacy_instance_id,
> > >    and so instead of using the ID you'd naturally get as a
> > >    written-from-scratch qdev device, it uses the legacy value
> > >    you pass in
> >
> > no, it only sets the SaveStateEntry.alias_id, which is only used
> > during incoming migration in find_se().
> >
> > Iow, it only works old->new.
> >
> > >  * thus the device/board in both old and new versions of QEMU
> > >    uses the same value and migration in both directions works
> >
> > sadly no
> >
> > >
> > > I don't understand why we would ever be using a "new 0-based
> > > instance_id" -- it seems to me that the whole point of setting
> > > a legacy ID value is that we will use it always, and I don't
> > > understand how the board code can know that it's going to be
> > > the target of an old->new migration as opposed to being the
> > > source of a new->old migration such that it can end up with
> > > a different ID value in the latter case.
> >
> > The target will find the "legacy" alias with find_se() on incoming
> > migration, but any new outgoing migration will use the new 0-based
> > instance_id
> >
> > >
> > > If qdev_set_legacy_instance_id() doesn't work the way I
> > > think it does above, what *does* it do ?
> >
> > just set the old alias_id for incoming migration.
> >
> > David, is that correct?
>
> Yes, I think it is.
> However, I'm curious which devices you're finding are explicitly setting
> their id's;  there aren't many - although there are some that probably
> should!
> For example, running an x86 image with:
>    -device isa-parallel,chardev=... -device isa-serial -device isa-serial -trace enable=qemu_loadvm_state_section_startfull
>
> shows:
> qemu_loadvm_state_section_startfull(uint32_t section_id, const char *idstr, uint32_t instance_id, uint32_t version_id) "%u(%s) %u %u"
>
> 165217@1576179638.856300:qemu_loadvm_state_section_startfull 41(serial) 0 3
> 165217@1576179638.856307:qemu_loadvm_state_section_startfull 42(serial) 1 3
> 165217@1576179638.856311:qemu_loadvm_state_section_startfull 43(parallel_isa) 0 1
>
> so those two serial devices are instances '0' and '1' I think by luck of
> their command line order, rather than having specified their base
> address (which would have been safer).

None of the qdev devices using vmsd use a hard-coded instance id
afaik. In device_set_realized(), vmstate_register_with_alias_id() is
called with instance_id = -1, so it relies on
calculate_new_instance_id(se->idstr)

>
> Dave
>
>
>
> > thanks
> >
> >
> > --
> > Marc-André Lureau
> >
> --
> Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
>


-- 
Marc-André Lureau

Re: [PATCH v4 20/37] mips: use sysbus_mmio_get_region() instead of internal fields

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Register the memory region with sysbus_init_mmio() and look it up with
> sysbus_mmio_get_region() to avoid accessing internal device fields.
> 
> Suggested-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/char/serial.c       | 1 +
>   hw/mips/mips_mipssim.c | 3 ++-
>   2 files changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 23f0b02516..02c545ff8c 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -993,6 +993,7 @@ static void serial_io_realize(DeviceState *dev, Error **errp)
>       qdev_init_nofail(DEVICE(s));
>   
>       memory_region_init_io(&s->io, NULL, &serial_io_ops, s, "serial", 8);
> +    sysbus_init_mmio(SYS_BUS_DEVICE(sio), &s->io);
>       sysbus_init_irq(SYS_BUS_DEVICE(sio), &s->irq);
>   }
>   
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 2c2c7f25b2..84c03dd035 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -227,7 +227,8 @@ mips_mipssim_init(MachineState *machine)
>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>           qdev_init_nofail(dev);
>           sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> -        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->serial.io);
> +        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8,
> +                      sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0));

TIL sysbus_add_io(), nice.
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>       }
>   
>       if (nd_table[0].used)
> 

Re: [PATCH v4 19/37] mips: use sysbus_add_io()

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/mips/mips_mipssim.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/hw/mips/mips_mipssim.c b/hw/mips/mips_mipssim.c
> index 3cd0e6eb33..2c2c7f25b2 100644
> --- a/hw/mips/mips_mipssim.c
> +++ b/hw/mips/mips_mipssim.c
> @@ -227,8 +227,7 @@ mips_mipssim_init(MachineState *machine)
>           qdev_set_legacy_instance_id(dev, 0x3f8, 2);
>           qdev_init_nofail(dev);
>           sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, env->irq[4]);
> -        memory_region_add_subregion(get_system_io(), 0x3f8,
> -                                    &SERIAL_IO(dev)->serial.io);
> +        sysbus_add_io(SYS_BUS_DEVICE(dev), 0x3f8, &SERIAL_IO(dev)->serial.io);

I'd simply squash this with the next patch, anyway:
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

>       }
>   
>       if (nd_table[0].used)
> 

Re: [PATCH v4 32/37] omap-intc: remove PROP_PTR

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Since clocks are not QOM objects, replace PROP_PTR of clocks with
> setters methods.
> 
> (in theory there should probably be different methods for omap1 &
> omap2 intc, but this is left as a future improvement)
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/arm/omap1.c        |  4 ++--
>   hw/arm/omap2.c        |  4 ++--
>   hw/intc/omap_intc.c   | 17 ++++++++++-------
>   include/hw/arm/omap.h | 14 ++++++++++++++
>   4 files changed, 28 insertions(+), 11 deletions(-)
> 
> diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
> index 6ce038a453..1afd1d3d7f 100644
> --- a/hw/arm/omap1.c
> +++ b/hw/arm/omap1.c
> @@ -3889,7 +3889,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
>   
>       s->ih[0] = qdev_create(NULL, "omap-intc");
>       qdev_prop_set_uint32(s->ih[0], "size", 0x100);
> -    qdev_prop_set_ptr(s->ih[0], "clk", omap_findclk(s, "arminth_ck"));
> +    omap_intc_set_iclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "arminth_ck"));
>       qdev_init_nofail(s->ih[0]);
>       busdev = SYS_BUS_DEVICE(s->ih[0]);
>       sysbus_connect_irq(busdev, 0,
> @@ -3899,7 +3899,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
>       sysbus_mmio_map(busdev, 0, 0xfffecb00);
>       s->ih[1] = qdev_create(NULL, "omap-intc");
>       qdev_prop_set_uint32(s->ih[1], "size", 0x800);
> -    qdev_prop_set_ptr(s->ih[1], "clk", omap_findclk(s, "arminth_ck"));
> +    omap_intc_set_iclk(OMAP_INTC(s->ih[1]), omap_findclk(s, "arminth_ck"));
>       qdev_init_nofail(s->ih[1]);
>       busdev = SYS_BUS_DEVICE(s->ih[1]);
>       sysbus_connect_irq(busdev, 0,
> diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
> index 457f152bac..1d7cc435ef 100644
> --- a/hw/arm/omap2.c
> +++ b/hw/arm/omap2.c
> @@ -2308,8 +2308,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
>       /* Actually mapped at any 2K boundary in the ARM11 private-peripheral if */
>       s->ih[0] = qdev_create(NULL, "omap2-intc");
>       qdev_prop_set_uint8(s->ih[0], "revision", 0x21);
> -    qdev_prop_set_ptr(s->ih[0], "fclk", omap_findclk(s, "mpu_intc_fclk"));
> -    qdev_prop_set_ptr(s->ih[0], "iclk", omap_findclk(s, "mpu_intc_iclk"));
> +    omap_intc_set_fclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "mpu_intc_fclk"));
> +    omap_intc_set_iclk(OMAP_INTC(s->ih[0]), omap_findclk(s, "mpu_intc_iclk"));
>       qdev_init_nofail(s->ih[0]);
>       busdev = SYS_BUS_DEVICE(s->ih[0]);
>       sysbus_connect_irq(busdev, 0,
> diff --git a/hw/intc/omap_intc.c b/hw/intc/omap_intc.c
> index 854b709ca0..73bb1c2af4 100644
> --- a/hw/intc/omap_intc.c
> +++ b/hw/intc/omap_intc.c
> @@ -38,10 +38,6 @@ struct omap_intr_handler_bank_s {
>       unsigned char priority[32];
>   };
>   
> -#define TYPE_OMAP_INTC "common-omap-intc"
> -#define OMAP_INTC(obj) \
> -    OBJECT_CHECK(struct omap_intr_handler_s, (obj), TYPE_OMAP_INTC)
> -
>   struct omap_intr_handler_s {
>       SysBusDevice parent_obj;
>   
> @@ -391,9 +387,18 @@ static void omap_intc_realize(DeviceState *dev, Error **errp)
>       }
>   }
>   
> +void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk)
> +{
> +    intc->iclk = clk;
> +}
> +
> +void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk)
> +{
> +    intc->fclk = clk;
> +}
> +
>   static Property omap_intc_properties[] = {
>       DEFINE_PROP_UINT32("size", struct omap_intr_handler_s, size, 0x100),
> -    DEFINE_PROP_PTR("clk", struct omap_intr_handler_s, iclk),
>       DEFINE_PROP_END_OF_LIST(),
>   };
>   
> @@ -647,8 +652,6 @@ static void omap2_intc_realize(DeviceState *dev, Error **errp)
>   static Property omap2_intc_properties[] = {
>       DEFINE_PROP_UINT8("revision", struct omap_intr_handler_s,
>       revision, 0x21),
> -    DEFINE_PROP_PTR("iclk", struct omap_intr_handler_s, iclk),
> -    DEFINE_PROP_PTR("fclk", struct omap_intr_handler_s, fclk),
>       DEFINE_PROP_END_OF_LIST(),
>   };
>   
> diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
> index f3aa670036..bcecf19f89 100644
> --- a/include/hw/arm/omap.h
> +++ b/include/hw/arm/omap.h
> @@ -67,6 +67,20 @@ void omap_clk_setrate(omap_clk clk, int divide, int multiply);
>   int64_t omap_clk_getrate(omap_clk clk);
>   void omap_clk_reparent(omap_clk clk, omap_clk parent);
>   
> +/* omap_intc.c */
> +#define TYPE_OMAP_INTC "common-omap-intc"
> +#define OMAP_INTC(obj)                                              \
> +    OBJECT_CHECK(omap_intr_handler, (obj), TYPE_OMAP_INTC)
> +
> +typedef struct omap_intr_handler_s omap_intr_handler;
> +
> +/*
> + * TODO: Ideally we should have a clock framework that
> + * let us wire these clocks up with QOM properties or links.
> + */
> +void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
> +void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
> +
>   /* OMAP2 l4 Interconnect */
>   struct omap_l4_s;
>   struct omap_l4_region_s {
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 33/37] omap-i2c: remove PROP_PTR

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Since clocks are not QOM objects, replace PROP_PTR of clocks with
> setters methods.
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Reviewed-by: Corey Minyard <cminyard@mvista.com>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/arm/omap1.c        |  2 +-
>   hw/arm/omap2.c        |  8 ++++----
>   hw/i2c/omap_i2c.c     | 19 ++++++++++++-------
>   include/hw/arm/omap.h | 13 +++++++++++++
>   4 files changed, 30 insertions(+), 12 deletions(-)
> 
> diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
> index 1afd1d3d7f..807e5f70d1 100644
> --- a/hw/arm/omap1.c
> +++ b/hw/arm/omap1.c
> @@ -4030,7 +4030,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
>   
>       s->i2c[0] = qdev_create(NULL, "omap_i2c");
>       qdev_prop_set_uint8(s->i2c[0], "revision", 0x11);
> -    qdev_prop_set_ptr(s->i2c[0], "fclk", omap_findclk(s, "mpuper_ck"));
> +    omap_i2c_set_fclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "mpuper_ck"));
>       qdev_init_nofail(s->i2c[0]);
>       busdev = SYS_BUS_DEVICE(s->i2c[0]);
>       sysbus_connect_irq(busdev, 0, qdev_get_gpio_in(s->ih[1], OMAP_INT_I2C));
> diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
> index 1d7cc435ef..171e2d0472 100644
> --- a/hw/arm/omap2.c
> +++ b/hw/arm/omap2.c
> @@ -2425,8 +2425,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
>   
>       s->i2c[0] = qdev_create(NULL, "omap_i2c");
>       qdev_prop_set_uint8(s->i2c[0], "revision", 0x34);
> -    qdev_prop_set_ptr(s->i2c[0], "iclk", omap_findclk(s, "i2c1.iclk"));
> -    qdev_prop_set_ptr(s->i2c[0], "fclk", omap_findclk(s, "i2c1.fclk"));
> +    omap_i2c_set_iclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "i2c1.iclk"));
> +    omap_i2c_set_fclk(OMAP_I2C(s->i2c[0]), omap_findclk(s, "i2c1.fclk"));
>       qdev_init_nofail(s->i2c[0]);
>       busdev = SYS_BUS_DEVICE(s->i2c[0]);
>       sysbus_connect_irq(busdev, 0,
> @@ -2437,8 +2437,8 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
>   
>       s->i2c[1] = qdev_create(NULL, "omap_i2c");
>       qdev_prop_set_uint8(s->i2c[1], "revision", 0x34);
> -    qdev_prop_set_ptr(s->i2c[1], "iclk", omap_findclk(s, "i2c2.iclk"));
> -    qdev_prop_set_ptr(s->i2c[1], "fclk", omap_findclk(s, "i2c2.fclk"));
> +    omap_i2c_set_iclk(OMAP_I2C(s->i2c[1]), omap_findclk(s, "i2c2.iclk"));
> +    omap_i2c_set_fclk(OMAP_I2C(s->i2c[1]), omap_findclk(s, "i2c2.fclk"));
>       qdev_init_nofail(s->i2c[1]);
>       busdev = SYS_BUS_DEVICE(s->i2c[1]);
>       sysbus_connect_irq(busdev, 0,
> diff --git a/hw/i2c/omap_i2c.c b/hw/i2c/omap_i2c.c
> index 3ba965a58f..3ccbd5cc2c 100644
> --- a/hw/i2c/omap_i2c.c
> +++ b/hw/i2c/omap_i2c.c
> @@ -28,10 +28,7 @@
>   #include "qemu/error-report.h"
>   #include "qapi/error.h"
>   
> -#define TYPE_OMAP_I2C "omap_i2c"
> -#define OMAP_I2C(obj) OBJECT_CHECK(OMAPI2CState, (obj), TYPE_OMAP_I2C)
> -
> -typedef struct OMAPI2CState {
> +struct OMAPI2CState {
>       SysBusDevice parent_obj;
>   
>       MemoryRegion iomem;
> @@ -56,7 +53,7 @@ typedef struct OMAPI2CState {
>       uint8_t divider;
>       uint8_t times[2];
>       uint16_t test;
> -} OMAPI2CState;
> +};
>   
>   #define OMAP2_INTR_REV	0x34
>   #define OMAP2_GC_REV	0x34
> @@ -504,10 +501,18 @@ static void omap_i2c_realize(DeviceState *dev, Error **errp)
>       }
>   }
>   
> +void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk)
> +{
> +    i2c->iclk = clk;
> +}
> +
> +void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk)
> +{
> +    i2c->fclk = clk;
> +}
> +
>   static Property omap_i2c_properties[] = {
>       DEFINE_PROP_UINT8("revision", OMAPI2CState, revision, 0),
> -    DEFINE_PROP_PTR("iclk", OMAPI2CState, iclk),
> -    DEFINE_PROP_PTR("fclk", OMAPI2CState, fclk),
>       DEFINE_PROP_END_OF_LIST(),
>   };
>   
> diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
> index bcecf19f89..39a295ba20 100644
> --- a/include/hw/arm/omap.h
> +++ b/include/hw/arm/omap.h
> @@ -81,6 +81,19 @@ typedef struct omap_intr_handler_s omap_intr_handler;
>   void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
>   void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
>   
> +/* omap_i2c.c */
> +#define TYPE_OMAP_I2C "omap_i2c"
> +#define OMAP_I2C(obj) OBJECT_CHECK(OMAPI2CState, (obj), TYPE_OMAP_I2C)
> +
> +typedef struct OMAPI2CState OMAPI2CState;
> +
> +/*
> + * TODO: Ideally we should have a clock framework that
> + * let us wire these clocks up with QOM properties or links.
> + */
> +void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk);
> +void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk);
> +
>   /* OMAP2 l4 Interconnect */
>   struct omap_l4_s;
>   struct omap_l4_region_s {
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 34/37] omap-gpio: remove PROP_PTR

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> Since clocks are not QOM objects, replace PROP_PTR of clocks with
> setters methods.
> 
> Move/adapt the existing TODO comment about a clock framework.
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/arm/omap1.c        |  2 +-
>   hw/arm/omap2.c        | 13 +++++++------
>   hw/gpio/omap_gpio.c   | 42 +++++++++++++++---------------------------
>   include/hw/arm/omap.h | 33 +++++++++++++++++++++++++++++----
>   4 files changed, 52 insertions(+), 38 deletions(-)
> 
> diff --git a/hw/arm/omap1.c b/hw/arm/omap1.c
> index 807e5f70d1..761cc17ea9 100644
> --- a/hw/arm/omap1.c
> +++ b/hw/arm/omap1.c
> @@ -4012,7 +4012,7 @@ struct omap_mpu_state_s *omap310_mpu_init(MemoryRegion *dram,
>   
>       s->gpio = qdev_create(NULL, "omap-gpio");
>       qdev_prop_set_int32(s->gpio, "mpu_model", s->mpu_model);
> -    qdev_prop_set_ptr(s->gpio, "clk", omap_findclk(s, "arm_gpio_ck"));
> +    omap_gpio_set_clk(OMAP1_GPIO(s->gpio), omap_findclk(s, "arm_gpio_ck"));
>       qdev_init_nofail(s->gpio);
>       sysbus_connect_irq(SYS_BUS_DEVICE(s->gpio), 0,
>                          qdev_get_gpio_in(s->ih[0], OMAP_INT_GPIO_BANK1));
> diff --git a/hw/arm/omap2.c b/hw/arm/omap2.c
> index 171e2d0472..e1c11de5ce 100644
> --- a/hw/arm/omap2.c
> +++ b/hw/arm/omap2.c
> @@ -2449,13 +2449,14 @@ struct omap_mpu_state_s *omap2420_mpu_init(MemoryRegion *sdram,
>   
>       s->gpio = qdev_create(NULL, "omap2-gpio");
>       qdev_prop_set_int32(s->gpio, "mpu_model", s->mpu_model);
> -    qdev_prop_set_ptr(s->gpio, "iclk", omap_findclk(s, "gpio_iclk"));
> -    qdev_prop_set_ptr(s->gpio, "fclk0", omap_findclk(s, "gpio1_dbclk"));
> -    qdev_prop_set_ptr(s->gpio, "fclk1", omap_findclk(s, "gpio2_dbclk"));
> -    qdev_prop_set_ptr(s->gpio, "fclk2", omap_findclk(s, "gpio3_dbclk"));
> -    qdev_prop_set_ptr(s->gpio, "fclk3", omap_findclk(s, "gpio4_dbclk"));
> +    omap2_gpio_set_iclk(OMAP2_GPIO(s->gpio), omap_findclk(s, "gpio_iclk"));
> +    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 0, omap_findclk(s, "gpio1_dbclk"));
> +    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 1, omap_findclk(s, "gpio2_dbclk"));
> +    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 2, omap_findclk(s, "gpio3_dbclk"));
> +    omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 3, omap_findclk(s, "gpio4_dbclk"));
>       if (s->mpu_model == omap2430) {
> -        qdev_prop_set_ptr(s->gpio, "fclk4", omap_findclk(s, "gpio5_dbclk"));
> +        omap2_gpio_set_fclk(OMAP2_GPIO(s->gpio), 4,
> +                            omap_findclk(s, "gpio5_dbclk"));
>       }
>       qdev_init_nofail(s->gpio);
>       busdev = SYS_BUS_DEVICE(s->gpio);
> diff --git a/hw/gpio/omap_gpio.c b/hw/gpio/omap_gpio.c
> index 41e1aa798c..85c16897ae 100644
> --- a/hw/gpio/omap_gpio.c
> +++ b/hw/gpio/omap_gpio.c
> @@ -40,10 +40,6 @@ struct omap_gpio_s {
>       uint16_t pins;
>   };
>   
> -#define TYPE_OMAP1_GPIO "omap-gpio"
> -#define OMAP1_GPIO(obj) \
> -    OBJECT_CHECK(struct omap_gpif_s, (obj), TYPE_OMAP1_GPIO)
> -
>   struct omap_gpif_s {
>       SysBusDevice parent_obj;
>   
> @@ -212,10 +208,6 @@ struct omap2_gpio_s {
>       uint8_t delay;
>   };
>   
> -#define TYPE_OMAP2_GPIO "omap2-gpio"
> -#define OMAP2_GPIO(obj) \
> -    OBJECT_CHECK(struct omap2_gpif_s, (obj), TYPE_OMAP2_GPIO)
> -
>   struct omap2_gpif_s {
>       SysBusDevice parent_obj;
>   
> @@ -747,21 +739,13 @@ static void omap2_gpio_realize(DeviceState *dev, Error **errp)
>       }
>   }
>   
> -/* Using qdev pointer properties for the clocks is not ideal.
> - * qdev should support a generic means of defining a 'port' with
> - * an arbitrary interface for connecting two devices. Then we
> - * could reframe the omap clock API in terms of clock ports,
> - * and get some type safety. For now the best qdev provides is
> - * passing an arbitrary pointer.
> - * (It's not possible to pass in the string which is the clock
> - * name, because this device does not have the necessary information
> - * (ie the struct omap_mpu_state_s*) to do the clockname to pointer
> - * translation.)
> - */
> +void omap_gpio_set_clk(omap_gpif *gpio, omap_clk clk)
> +{
> +    gpio->clk = clk;
> +}
>   
>   static Property omap_gpio_properties[] = {
>       DEFINE_PROP_INT32("mpu_model", struct omap_gpif_s, mpu_model, 0),
> -    DEFINE_PROP_PTR("clk", struct omap_gpif_s, clk),
>       DEFINE_PROP_END_OF_LIST(),
>   };
>   
> @@ -784,15 +768,19 @@ static const TypeInfo omap_gpio_info = {
>       .class_init    = omap_gpio_class_init,
>   };
>   
> +void omap2_gpio_set_iclk(omap2_gpif *gpio, omap_clk clk)
> +{
> +    gpio->iclk = clk;
> +}
> +
> +void omap2_gpio_set_fclk(omap2_gpif *gpio, uint8_t i, omap_clk clk)
> +{
> +    assert(i <= 5);
> +    gpio->fclk[i] = clk;
> +}
> +
>   static Property omap2_gpio_properties[] = {
>       DEFINE_PROP_INT32("mpu_model", struct omap2_gpif_s, mpu_model, 0),
> -    DEFINE_PROP_PTR("iclk", struct omap2_gpif_s, iclk),
> -    DEFINE_PROP_PTR("fclk0", struct omap2_gpif_s, fclk[0]),
> -    DEFINE_PROP_PTR("fclk1", struct omap2_gpif_s, fclk[1]),
> -    DEFINE_PROP_PTR("fclk2", struct omap2_gpif_s, fclk[2]),
> -    DEFINE_PROP_PTR("fclk3", struct omap2_gpif_s, fclk[3]),
> -    DEFINE_PROP_PTR("fclk4", struct omap2_gpif_s, fclk[4]),
> -    DEFINE_PROP_PTR("fclk5", struct omap2_gpif_s, fclk[5]),
>       DEFINE_PROP_END_OF_LIST(),
>   };
>   
> diff --git a/include/hw/arm/omap.h b/include/hw/arm/omap.h
> index 39a295ba20..6be386d0e2 100644
> --- a/include/hw/arm/omap.h
> +++ b/include/hw/arm/omap.h
> @@ -77,6 +77,16 @@ typedef struct omap_intr_handler_s omap_intr_handler;
>   /*
>    * TODO: Ideally we should have a clock framework that
>    * let us wire these clocks up with QOM properties or links.
> + *
> + * qdev should support a generic means of defining a 'port' with
> + * an arbitrary interface for connecting two devices. Then we
> + * could reframe the omap clock API in terms of clock ports,
> + * and get some type safety. For now the best qdev provides is
> + * passing an arbitrary pointer.
> + * (It's not possible to pass in the string which is the clock
> + * name, because this device does not have the necessary information
> + * (ie the struct omap_mpu_state_s*) to do the clockname to pointer
> + * translation.)
>    */
>   void omap_intc_set_iclk(omap_intr_handler *intc, omap_clk clk);
>   void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
> @@ -87,13 +97,28 @@ void omap_intc_set_fclk(omap_intr_handler *intc, omap_clk clk);
>   
>   typedef struct OMAPI2CState OMAPI2CState;
>   
> -/*
> - * TODO: Ideally we should have a clock framework that
> - * let us wire these clocks up with QOM properties or links.
> - */
> +/* TODO: clock framework (see above) */
>   void omap_i2c_set_iclk(OMAPI2CState *i2c, omap_clk clk);
>   void omap_i2c_set_fclk(OMAPI2CState *i2c, omap_clk clk);
>   
> +/* omap_gpio.c */
> +#define TYPE_OMAP1_GPIO "omap-gpio"
> +#define OMAP1_GPIO(obj)                                         \
> +    OBJECT_CHECK(struct omap_gpif_s, (obj), TYPE_OMAP1_GPIO)
> +
> +#define TYPE_OMAP2_GPIO "omap2-gpio"
> +#define OMAP2_GPIO(obj)                                         \
> +    OBJECT_CHECK(struct omap2_gpif_s, (obj), TYPE_OMAP2_GPIO)
> +
> +typedef struct omap_gpif_s omap_gpif;
> +typedef struct omap2_gpif_s omap2_gpif;
> +
> +/* TODO: clock framework (see above) */
> +void omap_gpio_set_clk(omap_gpif *gpio, omap_clk clk);
> +
> +void omap2_gpio_set_iclk(omap2_gpif *gpio, omap_clk clk);
> +void omap2_gpio_set_fclk(omap2_gpif *gpio, uint8_t i, omap_clk clk);
> +
>   /* OMAP2 l4 Interconnect */
>   struct omap_l4_s;
>   struct omap_l4_region_s {
> 

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 29/37] RFC: mips/cps: fix setting saar property

[Philippe Mathieu-Daudé]

On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> There is no "saar" property. Note: I haven't been able to test this
> code. Help welcome.
> 
> May fix commit 043715d1e0fbb3e3411be3f898c5b77b7f90327a ("target/mips:
> Update ITU to utilize SAARI and SAAR CP0 registers")
> 
> Cc: Aleksandar Markovic <amarkovic@wavecomp.com>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>   hw/mips/cps.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/hw/mips/cps.c b/hw/mips/cps.c
> index 1660f86908..c49868d5da 100644
> --- a/hw/mips/cps.c
> +++ b/hw/mips/cps.c
> @@ -106,7 +106,7 @@ static void mips_cps_realize(DeviceState *dev, Error **errp)
>           object_property_set_bool(OBJECT(&s->itu), saar_present, "saar-present",
>                                    &err);
>           if (saar_present) {
> -            qdev_prop_set_ptr(DEVICE(&s->itu), "saar", (void *)&env->CP0_SAAR);
> +            s->itu.saar = &env->CP0_SAAR;

Hmmm this could be what the author of 043715d1e wanted to do indeed.

Since this feature is incomplete (see comments in previous versions of 
this series:
   $ git grep saarp
   hw/mips/cps.c:98:    saar_present = (bool)env->saarp;
   target/mips/cpu.h:1103:    int saarp;
and I don't know how to test this, I'll defer to Aleksandar regarding 
this patch.

>           }
>           object_property_set_bool(OBJECT(&s->itu), true, "realized", &err);
>           if (err != NULL) {
> 

Re: [PATCH v4 27/37] leon3: use qdev gpio facilities for the PIL

[Philippe Mathieu-Daudé]

On 11/21/19 2:51 PM, Peter Maydell wrote:
> On Wed, 20 Nov 2019 at 15:30, Marc-André Lureau
> <marcandre.lureau@redhat.com> wrote:
>>
>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>> ---
>>   hw/sparc/leon3.c   | 6 ++++--
>>   target/sparc/cpu.h | 1 -
>>   2 files changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/hw/sparc/leon3.c b/hw/sparc/leon3.c
>> index cac987373e..1a89d44e57 100644
>> --- a/hw/sparc/leon3.c
>> +++ b/hw/sparc/leon3.c
>> @@ -230,8 +230,10 @@ static void leon3_generic_hw_init(MachineState *machine)
>>
>>       /* Allocate IRQ manager */
>>       dev = qdev_create(NULL, TYPE_GRLIB_IRQMP);
>> -    env->pil_irq = qemu_allocate_irq(leon3_set_pil_in, env, 0);
>> -    qdev_connect_gpio_out_named(dev, "grlib-irq", 0, env->pil_irq);
>> +    qdev_init_gpio_in_named_with_opaque(DEVICE(env), leon3_set_pil_in,
>> +                                        env, "pil", 1);
>> +    qdev_connect_gpio_out_named(dev, "grlib-irq", 0,
>> +                                qdev_get_gpio_in_named(DEVICE(env), "pil", 0));
>>       qdev_init_nofail(dev);
>>       sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, LEON3_IRQMP_OFFSET);
>>       env->irq_manager = dev;
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> 
> Creating a gpio pin on some object that isn't yourself
> looks a bit odd, but all this leon3 code is modifying
> the CPU object from the outside anyway. Someday we might
> tidy it up, but not today.

Watch out, while SPARCCPU inherits DeviceState from CPUState,
env does not: it is not QDEV but a CPUSPARCState object.

If you replace both DEVICE(env) uses by DEVICE(cpu), your patch looks safe.

If you agree with the change:
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Philippe Mathieu-Daudé]

Hi Marc-André,

On 12/11/19 1:01 PM, Marc-André Lureau wrote:
> Hi
> 
> On Sun, Dec 1, 2019 at 2:19 PM Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
>>
>>
>> - "chardev: generate an internal id when none given"
>>
>> As explained, this is necessary for qdev_prop_set_chr()
> 
> ping
> 
>>
>> - "serial: register vmsd with DeviceClass"
>>
>> This is standard qdev-ification, however it breaks backward migration,
>> but that's just how qdev_set_legacy_instance_id() works.
> 
> See thread, someone could review or nack (if backward migration is a problem).
> 
>>
>> - "sm501: make SerialMM a child, export chardev property"
>>
>> review?
> 
> ping

This would be simpler you include "hw/display/sm501: Always map the 
UART0" previous your changes.

To finish the serial QOM-ification, we need to make serial_reset a 
DeviceReset, then we can remove the qemu_register_reset() call.

>>
>> - "qdev/qom: remove some TODO limitations now that PROP_PTR is gone"
>>
>> This should be straightforward.
> 
> ping

Re: [EXTERNAL]Re: [PATCH v4 29/37] RFC: mips/cps: fix setting saar property

[Aleksandar Markovic]

> From: Philippe Mathieu-Daudé <philmd@redhat.com>
> Sent: Sunday, December 15, 2019 6:56 AM
> > On 11/20/19 4:24 PM, Marc-André Lureau wrote:
> > There is no "saar" property. Note: I haven't been able to test this
> > code. Help welcome.
> >
> > May fix commit 043715d1e0fbb3e3411be3f898c5b77b7f90327a ("target/mips:
> > Update ITU to utilize SAARI and SAAR CP0 registers")
> >
> > Cc: Aleksandar Markovic <amarkovic@wavecomp.com>
> > Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> > ---
> >   hw/mips/cps.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/hw/mips/cps.c b/hw/mips/cps.c
> > index 1660f86908..c49868d5da 100644
> > --- a/hw/mips/cps.c
> > +++ b/hw/mips/cps.c
> > @@ -106,7 +106,7 @@ static void mips_cps_realize(DeviceState *dev, Error **errp)
> >           object_property_set_bool(OBJECT(&s->itu), saar_present, "saar-present",
> >                                    &err);
> >           if (saar_present) {
> > -            qdev_prop_set_ptr(DEVICE(&s->itu), "saar", (void *)&env->CP0_SAAR);
> > +            s->itu.saar = &env->CP0_SAAR;
>
> Hmmm this could be what the author of 043715d1e wanted to do indeed.
> 
> Since this feature is incomplete (see comments in previous versions of
> this series:
>    $ git grep saarp
>    hw/mips/cps.c:98:    saar_present = (bool)env->saarp;
>    target/mips/cpu.h:1103:    int saarp;
> and I don't know how to test this, I'll defer to Aleksandar regarding
> this patch.

Hello, Philippe,

Good that you brought this to my attention - but Marc-André and
I already had a discussion about this in this series' cover letter
responses (unfortunately not followed up by me as a response to
this patch, sorry about that):

https://lists.gnu.org/archive/html/qemu-devel/2019-12/msg00056.html

In essence, our conclusion was that the real fix would be certainly
outside of the scope of this series, since it requires some really
(non-straightforward) mips-specific code changes, so we agreed
that Marc-André submit this patch as-is, and later on (certainly
within timeframe of 5.0) I will submit the fix addressing the root
cause (absence of "saar" property, and incomplete handling of
SAAR and SAARI config registers).

In other words, this patch is fine at this moment, and formally:

Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>

(Marc-André, just please remove "RFC" from the title, and remove
the sentence "Note: I haven't been able to test this code. Help
welcome." from the commit message, since we have the deal on
how and who will fix the underlining problem.)

Thanks to both of you!

Aleksandar

Re: [PATCH v4 27/37] leon3: use qdev gpio facilities for the PIL

[Marc-André Lureau]

On Sun, Dec 15, 2019 at 10:10 AM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/21/19 2:51 PM, Peter Maydell wrote:
> > On Wed, 20 Nov 2019 at 15:30, Marc-André Lureau
> > <marcandre.lureau@redhat.com> wrote:
> >>
> >> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> >> ---
> >>   hw/sparc/leon3.c   | 6 ++++--
> >>   target/sparc/cpu.h | 1 -
> >>   2 files changed, 4 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/hw/sparc/leon3.c b/hw/sparc/leon3.c
> >> index cac987373e..1a89d44e57 100644
> >> --- a/hw/sparc/leon3.c
> >> +++ b/hw/sparc/leon3.c
> >> @@ -230,8 +230,10 @@ static void leon3_generic_hw_init(MachineState *machine)
> >>
> >>       /* Allocate IRQ manager */
> >>       dev = qdev_create(NULL, TYPE_GRLIB_IRQMP);
> >> -    env->pil_irq = qemu_allocate_irq(leon3_set_pil_in, env, 0);
> >> -    qdev_connect_gpio_out_named(dev, "grlib-irq", 0, env->pil_irq);
> >> +    qdev_init_gpio_in_named_with_opaque(DEVICE(env), leon3_set_pil_in,
> >> +                                        env, "pil", 1);
> >> +    qdev_connect_gpio_out_named(dev, "grlib-irq", 0,
> >> +                                qdev_get_gpio_in_named(DEVICE(env), "pil", 0));
> >>       qdev_init_nofail(dev);
> >>       sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, LEON3_IRQMP_OFFSET);
> >>       env->irq_manager = dev;
> >
> > Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> >
> > Creating a gpio pin on some object that isn't yourself
> > looks a bit odd, but all this leon3 code is modifying
> > the CPU object from the outside anyway. Someday we might
> > tidy it up, but not today.
>
> Watch out, while SPARCCPU inherits DeviceState from CPUState,
> env does not: it is not QDEV but a CPUSPARCState object.
>
> If you replace both DEVICE(env) uses by DEVICE(cpu), your patch looks safe.
>
> If you agree with the change:
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

good catch (it was actually already fixed in my branch ;)
thanks

Re: [EXTERNAL]Re: [PATCH v4 29/37] RFC: mips/cps: fix setting saar property

[Philippe Mathieu-Daudé]

On 12/16/19 8:36 PM, Aleksandar Markovic wrote:
>> From: Philippe Mathieu-Daudé <philmd@redhat.com>
>> Sent: Sunday, December 15, 2019 6:56 AM
>>> On 11/20/19 4:24 PM, Marc-André Lureau wrote:
>>> There is no "saar" property. Note: I haven't been able to test this
>>> code. Help welcome.
>>>
>>> May fix commit 043715d1e0fbb3e3411be3f898c5b77b7f90327a ("target/mips:
>>> Update ITU to utilize SAARI and SAAR CP0 registers")
>>>
>>> Cc: Aleksandar Markovic <amarkovic@wavecomp.com>
>>> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
>>> ---
>>>    hw/mips/cps.c | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/hw/mips/cps.c b/hw/mips/cps.c
>>> index 1660f86908..c49868d5da 100644
>>> --- a/hw/mips/cps.c
>>> +++ b/hw/mips/cps.c
>>> @@ -106,7 +106,7 @@ static void mips_cps_realize(DeviceState *dev, Error **errp)
>>>            object_property_set_bool(OBJECT(&s->itu), saar_present, "saar-present",
>>>                                     &err);
>>>            if (saar_present) {
>>> -            qdev_prop_set_ptr(DEVICE(&s->itu), "saar", (void *)&env->CP0_SAAR);
>>> +            s->itu.saar = &env->CP0_SAAR;
>>
>> Hmmm this could be what the author of 043715d1e wanted to do indeed.
>>
>> Since this feature is incomplete (see comments in previous versions of
>> this series:
>>     $ git grep saarp
>>     hw/mips/cps.c:98:    saar_present = (bool)env->saarp;
>>     target/mips/cpu.h:1103:    int saarp;
>> and I don't know how to test this, I'll defer to Aleksandar regarding
>> this patch.
> 
> Hello, Philippe,
> 
> Good that you brought this to my attention - but Marc-André and
> I already had a discussion about this in this series' cover letter
> responses (unfortunately not followed up by me as a response to
> this patch, sorry about that):
> 
> https://lists.gnu.org/archive/html/qemu-devel/2019-12/msg00056.html
> 
> In essence, our conclusion was that the real fix would be certainly
> outside of the scope of this series, since it requires some really
> (non-straightforward) mips-specific code changes, so we agreed
> that Marc-André submit this patch as-is, and later on (certainly
> within timeframe of 5.0) I will submit the fix addressing the root
> cause (absence of "saar" property, and incomplete handling of
> SAAR and SAARI config registers).

OK, thanks for clarifying again!

> In other words, this patch is fine at this moment, and formally:
> 
> Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>
> 
> (Marc-André, just please remove "RFC" from the title, and remove
> the sentence "Note: I haven't been able to test this code. Help
> welcome." from the commit message, since we have the deal on
> how and who will fix the underlining problem.)
> 
> Thanks to both of you!
> 
> Aleksandar
> 

Re: [PATCH v4 00/37] Clean-ups: qom-ify serial and remove QDEV_PROP_PTR

[Marc-André Lureau]

Hi

Still trying to make progress on this series (which is preliminary to
other pending work..):

On Wed, Dec 11, 2019 at 4:01 PM Marc-André Lureau
<marcandre.lureau@gmail.com> wrote:
>
> Hi
>
> On Sun, Dec 1, 2019 at 2:19 PM Marc-André Lureau
> <marcandre.lureau@gmail.com> wrote:
> >
> >
> > - "chardev: generate an internal id when none given"
> >
> > As explained, this is necessary for qdev_prop_set_chr()
>
> ping
>

ping

> >
> > - "serial: register vmsd with DeviceClass"
> >
> > This is standard qdev-ification, however it breaks backward migration,
> > but that's just how qdev_set_legacy_instance_id() works.
>
> See thread, someone could review or nack (if backward migration is a problem).
>

ping

> >
> > - "sm501: make SerialMM a child, export chardev property"
> >
> > review?
>
> ping
>

rebased on top of "hw/display/sm501: Always map the UART0" as
requested by Philippe. Can wait for v5 to get review.


> >
> > - "qdev/qom: remove some TODO limitations now that PROP_PTR is gone"
> >
> > This should be straightforward.
>
> ping

ping


thanks


-- 
Marc-André Lureau